2.1. Sodium Borohydride As a Fuel: Direct Borohydride Fuel Cell

University of Southampton Research Repository ePrints Soton

Copyright © and Moral Rights for this thesis are retained by the author and/or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This thesis cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder/s. The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.

When referring to this work, full bibliographic details including the author, title, awarding institution and date of the thesis must be given e.g.

AUTHOR (year of submission) "Full thesis title", University of Southampton, name of the University School or Department, PhD Thesis, pagination

http://eprints.soton.ac.uk

UNIVERSITY OF SOUTHAMPTON

FACULTY OF ENGINEERING AND THE ENVIRONMENT

Engineering Sciences

Investigation of the use of sodium borohydride for fuel cells

Irene Merino Jimenez

Thesis for the degree of Doctor of Philosophy

November 2013

UNIVERSITY OF SOUTHAMPTON ABSTRACT FACULTY OF ENGINEERING AND THE ENVIRONMENT ENGINEERING SCIENCES Doctor of Philosophy INVESTIGATION OF THE USE OF SODIUM BOROHYDRIDE FOR FUEL CELLS Irene Merino Jimenez

The use of NaBH4 for fuel cells offers a promising alternative to incumbent electrical power generation technologies. The predicted high energy density (9.3 kW h kg-1) of the direct borohydride fuel cell (DBFC) and its capacity to release 8e- per molecule converts it to a potential substitute for the H2/O2 system. Sodium borohydride, with 10.6 wt. % hydrogen content, can also generate H2 gas to be fed into a traditional H2/O2 fuel cell in an indirect borohydride fuel cell (IBFC). However, there are fundamental aspects of the DBFC and the IBFC that need to be addressed to achieve their optimal performance. The hydrolysis of borohydride is the key factor in both cases, being the undesired parallel reaction that occurs during the borohydride oxidation in the DBFC, and the main reaction taking place and to be promoted in the IBFC. The competition - between BH4 oxidation and its hydrolysis depends on the electrode material, electrolyte composition and operation conditions, such as the temperature.

In this work, two approaches to the use of NaBH4 for fuel cells are considered. Catalysts such as Pd-Ir alloy on microbifrous carbon, gold coated reticulated vitreous carbon (RVC), planar gold and dispersed nanoparticulate gold supported on carbon (Au/C), were tested for the direct borohydride oxidation while Pd-Ir alloy, Pt nanoparticles on carbon paper and Pd deposited on granular carbon (Pd/C) were evaluated to generate H2 from NaBH4 for use in an IBFC. The gold coated RVC electrodes demonstrated good activity towards the borohydride oxidation, increasing the kinetic rate constants and the current density with the thickness of the coating and the porosity grades. The Pd-Ir alloy was also catalytic towards the DBFC, with current densities between 100 and 200 mA -2 3 -1 cm but with low H2 generation rates (< 0.1 cm min ). As computational methods could play a prominent role in the design and characterisation of DBFCs, density functional theory (DFT) was used to investigate the reaction mechanism of borohydride i

oxidation at Pd-Ir surfaces. This work also studies the use of surfactants, including Triton X-100, Zonyl FSO, S-228M, sodium dodecyl sulphate and FC4430, during the direct oxidation of borohydride ions using a planar gold and Au/C electrodes. The addition of 0.001 wt. % Triton X-100 to the alkaline borohydride solution decreased the

H2 generation by 23 % at the Au/C electrode, while the borohydride oxidation remained unaffected. In contrast, in the H2 generator, the Pd/C catalyst showed an excellent activity towards the borohydride hydrolysis, obtaining a maximum rate of 8 × 103 cm3 -1 -1 -3 3 min gmetal during 120 minutes using 4 mol dm NaBH4 in 350 cm distilled water and

15 g of catalyst. This is the highest H2 generation rate reported in a laboratory scale reactor using borohydride and a Pd base catalyst.

Keywords: borohydride oxidation and decomposition, catalysis, DBFC, IBFC, hydrogen generation, hydrolysis inhibition, kinetic rate constant.

Contents

ABSTRACT ...... i

Contents ...... iii

List of tables ...... vi

List of figures ...... vii

DECLARATION OF AUTHORSHIP ...... xvii

Acknowledgements ...... xix

Abbreviations ...... xxi

Symbols ...... xxii

Chapter 1: Introduction ...... 1

1.1. Background and motivation ...... 1 1.2. Aims and objectives ...... 8 1.3. Thesis outline ...... 9 Chapter 2: Literature Review ...... 11

2.1. Sodium borohydride as a fuel: Direct Borohydride Fuel Cell ...... 12 2.1.1. Anode ...... 12 2.1.2. Cathode ...... 33 2.1.3. Membranes ...... 36 2.1.4. Cell performance ...... 40 2.1.5. Hydrogen evolution and the use of inhibitors ...... 42 2.1.6. Influence of operational conditions in a DBFC ...... 49 2.1.7. Engineering aspects of direct borohydride fuel cells ...... 53 2.1.8. Modelling and Simulation ...... 58 2.1.9. Recycling sodium metaborate product to sodium borohydride reactant...... 62 2.1.10. Synthesis of sodium borohydride ...... 65 2.2. Sodium borohydride for hydrogen generation ...... 66 2.2.1. Hydrogen generation from metal hydrides – Sodium borohydride ...... 66 2.2.2. Engineering aspects of the indirect borohydride fuel cell ...... 74 2.3. Summary ...... 79 Chapter 3: Experimental methodology ...... 83

3.1. Materials and chemicals ...... 83 iii

3.2. Equipment ...... 84 3.3. Electrochemical cell and methodology ...... 84 3.3.1. Electrolysis ...... 84 - 3.3.2. Effects of surfactants on the BH4 oxidation kinetics ...... 87 3.4. Borohydride oxidation at Pd-Ir/Ti electrode ...... 88 3.5. Gold coated RVC electrodes for borohydride oxidation ...... 90 3.5.1. PVD sputtering technique ...... 90 3.5.2. Cyclic voltammetry ...... 90 3.6. Hydrogen generator design and testing ...... 91 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation ...... 96

4.1. Electrolysis ...... 96 4.2. Cyclic voltammetry ...... 102 4.3. Computational methods...... 105 4.3.1. Pd-Ir(111) crystalline structures ...... 114 4.3.2. Activation energy ...... 116 4.4. DFT results and discussion ...... 117

4.4.1. Borohydride ion adsorption over Pd2-Ir1(111) and Pd2-Ir2(111) ...... 117 - 4.4.2. Elementary surface energies of BH4 oxidation at Pd2Ir1(111) and Pd2Ir2(111) surfaces ...... 120 4.4.3. Competitive borohydride oxidation versus hydrolysis at the Pd-Ir surfaces 130 4.5. Conclusions ...... 137 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation ... 138

5.1. Electrolysis at constant current ...... 138 5.1.1. Planar Au electrode ...... 140 5.1.2. Dispersed Au/C electrode...... 141 5.2. Potentiostatic electrolysis ...... 142 5.2.1. Planar Au electrode ...... 142 5.2.2. Au/C Electrode ...... 147 5.2.3. Triton X-100 ...... 155 5.2.4 Fuel utilization...... 165 5.3. The effect of surfactants on the kinetic parameters of borohydride oxidation .. 169 5.3.1. Planar Au electrode ...... 169 5.3.1.1. CV at a stationary electrode ...... 169 5.3.2. Au/C electrode...... 189 5.4. Conclusions ...... 194 iv

Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation ...... 196

6.1. Cyclic voltammetry ...... 197 6.2. Kinetic constants ...... 205 6.3. Conclusions ...... 208 Chapter 7: Hydrogen generation from sodium borohydride ...... 210

7.1. Generation of hydrogen using platinised titanium catalyst ...... 211 7.2 Generation of hydrogen using Pd-Ir catalyst supported on carbon fibres...... 212 7.3 Generation of hydrogen using nanoparticulate platinum on carbon ...... 214 7.4. Palladium 0.5 wt. % supported on granular carbon catalyst (Pd/C) ...... 215 7.4.1. Importance of the water source in the hydrogen generation: distilled water, drinking water and river water ...... 215 7.4.2. Effect of catalyst loading ...... 219 7.4.3. Effect of NaOH stabilizer on the hydrogen generation ...... 223 7.4.4. Effect of the reused catalyst ...... 224

7.4.4 Hydrogen generation: mix of solid NaBH4 powder and Pd/C catalyst followed by the addition of still drinking water ...... 227 7.4.5. Scanning electron microscopy (SEM) imaging of unused and used Pd/C catalyst ...... 229 7.5. Nitrogen adsorption on the Pd/C catalyst ...... 233 7.6. Conclusions ...... 235 Chapter 8: Conclusions and Further Work ...... 238

8.1. Direct borohydride fuel cells ...... 238 8.1.1. Anode materials ...... 239 8.1.2. Effect of surfactants in the borohydride oxidation and its hydrolysis ...... 241 8.2. Hydrogen generation from sodium borohydride ...... 244 8.2.1. Hydrogen generation ...... 244 8.3. Suggestions for Further Work ...... 245 8.3.1. Direct borohydride fuel cell ...... 245 8.3.2. Indirect borohydride fuel cell ...... 247 Appendix I. Classification and properties of the surfactants used ...... 249 References ...... 250

List of tables

1. 1 Theoretical energies of various fuels and oxidants, assuming complete oxidation of the fuel...... 6

- 5. 1 Hydrogen generation rate during constant current (600 mA) electrolysis of BH4 at a planar Au and an Au/C electrode (each 3 cm2). The solutions contained 1 -3 -3 mol dm NaBH4 in 3 mol dm NaOH and the following concentration of surfactants: a) 0 wt. %, b) 0.00001 wt. %, c) 0.0001 wt. %, d) 0.001 wt. %. The surfactants used were: SDS, Triton X-100, Zonyl FSO and S-228M...... 139 5. 2 Diffusion coefficient of borohydride ion in the absence and the presence of various concentrations of SDS, 0.1 wt. % Triton X-100 and 0.1 wt. % FC4430 at 23 oC……...... 182 5. 3 Kinetic rate constants for borohydride oxidation at controlled potentials in an -3 -3 o electrolyte containing 0.02 mol dm NaBH4 in 3 mol dm NaOH at 23 C with different concentrations of SDS from 0 wt. % to 0.1 wt. %...... 185

6. 1 Heterogeneous and kinetic rate constants for the different gold coated RVC electrodes at 298 K. The concentration of borohydride ions was 0.02 mol dm-3 -3 NaBH4 in 3 mol dm NaOH...... 207

7. 1 Comparison of the results obtained in this work and relevant literature...... 222 7. 2 Elemental composition of the area within the square section of an unused Pd/C catalyst shown in Figure 7.11.a)...... 230 7. 3 Elemental composition of the circle area in the inset SEM image of Figure 7.11.b) of an unused Pd/C catalyst...... 231 7. 4 Elemental composition within the large square area of the main SEM image of Figure 7.11.c) of a section of the three times used palladium catalyst on granulated carbon...... 232 7. 5 Percentage weight of each element found on the cross marks points shown on Figure 7.11.d) using EDX spectra analysis...... 233

List of figures

1. 1 Diagram of the comparison of the hydrogen content (wt. %) of different hydrides.3 1. 2 Arrangement of an indirect borohydride fuel cell (IBFC) system composed by a

hydrogen generator from NaBH4, which product feeds a H2/O2 fuel cell [1]...... 4 1. 3 Arrangement of a direct borohydride-hydrogen peroxide fuel flow cell system (DBFC) [23]...... 7

2. 1 Evolution of the number of publications related to the direct borohydride fuel cell (DBFC) and the hydrolysis of borohydride. The inset shows a close up between the years 2001 and 2012 [2]...... 11 2. 2 Reaction pathways involving anodic oxidation of borohydride ion and competitive reactions [35, 37]...... 16 - 2. 3 Power density (P) vs. current density in a BH4 /O2 DBFC with various anodes of 2 4 cm active area: a) Zr0.9Ti0.1Mn0.6V0.2Co0.1Ni1.1, b) Pt-Ni, c) Pd/C, d) Au/Ti, e) Au/C, f) Ag/Ti, g) Pt/C, h) Ni/C. Cathode: Pt/C (2 mg Pt cm-2). Membrane: Nafion 117. Temperature: 85oC, except of b) 60oC. Fuel: a): 10 wt. % (2.64 mol -3 -3 dm ) NaBH4 in 20 wt. % (5 mol dm ) NaOH [23] and b) to h): 5 wt. % (1.32 -3 -3 mol dm ) NaBH4 in 10 wt. % (2.5 mol dm ) NaOH [39, 40]...... 18

2. 4 Power density (P) vs. current density curves measured in a BH4/H2O2 fuel cell at 25 oC using a Pd/C anode of 25 cm2 active area and different cathode materials o [43]: a) Pd, b) Pt, c) Ag; and in a BH4/O2 fuel cell at 85 C, using an Au/C anode of 4 cm2 active area (2 mg Au cm2) and different cathode materials (2 mg cm2): d) Pt/C, e) FeTMPP, f) Ni/C and g) Ag/C. Membrane: Nafion 117.. Fuel: 5 wt. % 3 -3 (1.32 mol dm ) NaBH4 in 10 wt. % (2.5 mol dm ) NaOH [2, 77, 87]...... 36 2. 5 Different configurations of the direct borohydride fuel cell (DBFC): MEA configuration where the electrodes are separated by: a) an anionic exchange membrane and b) a cationic exchange membrane. Oxygen reduction in an alkaline media in a flow system where the electrodes are separated by the electrolytes and by c) an anionic exchange membrane and d) a cationic exchange membrane...... 38 2. 6 A parallel-plate electrode unit cell showing the location of elements that contributes to the overall cell voltage drop [2]...... 41 vii

2. 7 Schematic diagram of a parallel-plate electrode cell showing the elements that contributes to the overall cell voltage drop. The voltage across the electrodes, electrolyte and membrane is plotted versus the distance between the two parallel electrodes [94]...... 42 2. 8 Hydrogen evolution rate vs. anodic current. a) 1 g Ni (Inco type 210, particle size -3 0.5–1.0 μm ), d) 10 wt. % Pd/C, g) 5 wt. % Pt/C: 0.5 mol dm NaBH4 in 6 mol dm-3 NaOH at room temperature [31]; b) Ni (Inco type 255 particle size 2.2–2.8 -3 μm), c) 20 wt. % Pt/C e) 20 wt. % Au/C, f) Cu: 0.5 mol dm NaBH4 in 2 mol dm-3 NaOH [95]...... 43 2. 9 Cyclic voltammogram on colloidal 10 wt. % Os supported on carbon support -1 -3 - (Vulcan XC-72) at 100 mV s and 298 K [42]: 1) 0.03 mol dm BH4 + 2 mol dm-3 NaOH + 1.5 × 10-3 TU, 2) 2 mol dm-3 NaOH + 1.5 × 10-3 TU and 3) 2 mol -3 -3 -3 dm NaOH; together with CV of 0.03 mol dm NaBH4 + 2 mol dm NaOH on Pt dis electrode (1 mm diameter): 4) in the absence of TU and 5) in the presence of 1.5 × 10-3 mol dm-3 TU [32]...... 47 2. 10 Typical three-electrode glass cell to carry out fundamental voltammetric studies of the borohydride oxidation reaction such as kinetic rate constants, exchange current densities and mass transport characteristics [109]...... 53 2. 11 Different configurations of flow arrangements in a bipolar flow cell: a) 4-cell stack with the reactant flow circuit feed in parallel flow circuit and b) a 5-cell stack with the reactant feed in series...... 57 2. 12 Reaction mechanism of the hydrolysis of borohydride ions on a metal surface [135]…………...... 67 2. 13 Arrangement of an indirect borohydride fuel cell (IBFC) system composed by a

hydrogen generator from NaBH4, which product feeds a H2/O2 fuel cell [1]...... 75 2. 14 Improved design for hydrogen generation from sodium borohydride, including a heat exchanger inside the reactor for fuel preheating [105]...... 76

3. 1 Assembly used for hydrogen generation measurements during electrolysis of borohydride...... 85 3. 2 Photograph of the experimental arrangement with a gold rotating disk working electrode...... 87

viii

3. 3 Rotating disk electrode assemblies...... 88 3. 4 SEM image of A) carbon fibers deposited by direct charging electrostatic flocking. B) Pd-Ir coated carbon microfiber array [165]...... 89 3. 5 Three-electrode cell...... 90 3. 6 Three-dimensional schematic diagram of the proposed hydrogen generator reactor [166]...... 93 3. 7 Photograph of the built hydrogen generator reactor…………………………….94

4. 1 Hydrogen gas generation rate vs. potential applied at the Pd/Ir coated 2 -3 -3 microfibrous carbon electrode (3 cm ). a) 0.5 mol dm NaBH4 in 3 mol dm -3 -3 NaOH, b) 1 mol dm NaBH4 in 3 mol dm NaOH...... 98 4. 2 Gas generation rate vs. current from electrolysis at constant potential: a) Pd/Ir -3 coated microfibrous carbon electrode using a solution of 0.5 mol dm NaBH4 in 3 mol dm-3 NaOH, b) Pd/Ir coated microfibrous carbon electrode and 1 mol dm-3 -3 -3 NaBH4 in 3 mol dm NaOH, c) 10 wt. % Pd/C and 0.52 mol dm NaBH4 in 2 -3 -3 -3 mol dm NaOH. [31], d) 10 wt. % Pd/C and 1.02 mol dm NaBH4 in 2 mol dm NaOH [31]...... 99 4. 3 Gas generation rate vs. current obtained during experiments at constant potential, using different anode materials (3 cm2): a) Au flat plate electrode, b) Au/C electrode, c) Pd-Ir coated microfibrous carbon electrode using a solution -3 -3 -3 containing 1 mol dm NaBH4 in 3 mol dm NaOH. d) 0.5 mol dm NaBH4 in 2 mol dm-3 NaOH at Au/C (4 cm2 geometric area) (20 wt. %) [95]...... 101 4. 4 Apparent number of electrons released during oxidation vs. electrode potential -3 from the electrolysis of solutions containing: a) 0.5 mol dm NaBH4 in 3 mol -3 -3 -3 o dm NaOH, b) 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C using an Pd-Ir electrode...... 102 4.5 Cyclic voltammogram at the Pd-Ir coated microfibrous carbon electrode (3 cm2).

-3 -3 Solutions containing 0.02 mol dm NaBH4 in 3 mol dm NaOH and 3 mol dm-3 NaOH……………………………………………………………….103

4.6 Flow chart describing the computational procedure for the calculation of the total energy at the ground state……………………………………………………..114

- 4. 7 BH4 molecule adsorbed on: a) Pd2-Ir1(111) and b) Pd2-Ir2(111). The dark blue atoms correspond to Ir, the light blue slightly larger atoms correspond to Pd, the white atoms are H atoms and the pink atoms are B...... 117 - 4. 8 Free energy of adsorption of BH4 aq over Pd2Ir2(111) and Pd2Ir1(111) using - -3 vacuum slab model (T = 298 K, [BH4 aq] = 0.03 mol dm )...... 119 * * * 4. 9 BH3 dehydrogenation to BH2 +H over Pd2Ir1(111) surface: a) initial state, b) transition state, c) final state...... 122 * * * 4. 10 BH3 dehydrogenation to BH2 +H over Pd2Ir2(111) surface: a) initial state, b) transition state, c) final state...... 122

4. 11 Reaction energetics of the borohydride oxidation over Pd2Ir1(111) at -0.5 V vs. - -3 - -3 SHE, [BH4 ] = 0.02 mol dm , [OH ] = 2 mol dm , T = 298 K. The red line represents the most favourable mechanism of borohydride oxidation according to the relative energies and activation barriers calculated from DFT. The broken and dashed lines represent other possible mechanisms of borohydride oxidation analysed. Numbers on the plot label the activation barrier of each elementary step……………...... 123

4. 12 Reaction energetics of the borohydride oxidation over Pd2Ir2(111) at -0.5 V vs. - -3 - -3 SHE, [BH4 ] = 0.02 mol dm , [OH ] = 2 mol dm , T = 298 K. The red line represents the most favourable mechanism of borohydride oxidation according to the relative energies and activation barriers calculated from DFT. The broken and dashed lines represent other possible mechanisms of borohydride oxidation analysed. Numbers on the plot label the activation barrier of each elementary step……...... 127

4. 13 Configurations of the preferred borohydride oxidation path on Pd2Ir2(111) surface. The dark blue atoms correspond to Ir, the light blue big atoms correspond to Pd, the pink atoms are B, the red are O atoms and the white are H atoms…………...... 129

4. 14 Preferred reaction mechanism for the borohydride oxidation on Pd2Ir1 at -1 V vs. Hg/HgO and -0.64 V vs. HgO...... 130

4. 15 Preferred H atom adsorption over: a) Pd(111), b) Pd2Ir1(111) c) Pd2Ir1(111) surfaces at low coverage. The dark blue atoms correspond to Ir, the light blue big atoms correspond to Pd and the white atoms are H atoms...... 131

4. 16 Reaction free energies for H* oxidation (straight line) and evolution (dash line) of * H2 gas as a function of potential for low hydrogen coverage 1/9 ML H on * Pd2Ir1(111) surface (black lines), 1/16 ML H on Pd2Ir2(111) surface (red lines) * and 1/9 ML H on Pd(111) surface (blue lines). The Pd2Ir1(111) and the

Pd2Ir2(111) lines overlay each other...... 133

4. 17 H adsorption built for high H coverage over: a) Pd(111), b) Pd2Ir1(111) c)

Pd2Ir2(111) surfaces at high coverage. Preferred H adsorption optimized over: d)

Pd(111), e) Pd2Ir1(111) f) Pd2Ir2(111) surfaces at low coverage The dark blue atoms correspond to Ir, the light blue big atoms correspond to Pd and the white atoms are H atoms...... 135

5. 1 Gas generation rate vs. potential for a planar Au electrode. All the solutions -3 -3 o contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C and the presence of the surfactant was varied: a) without surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % SDS...... 143 5. 2 Current density vs. electrode potential at the planar Au anode. All the solutions -3 -3 o contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C. Some of them also contained surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430...... 145 5. 3 Hydrogen generation rate vs. current density at the planar Au anode. All the -3 -3 o solutions contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C. Some of them also contained surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS and d) 0.1 wt. % FC4430...... 146 5. 4 Hydrogen gas generation rate vs. current density in a solution containing 1 mol -3 -3 o dm NaBH4 in 3 mol dm NaOH at 23 C using an Au/C 10 wt. % electrode. 148 5. 5 a) Hydrogen gas generation rate vs. electrode potential and b) apparent number of electrons released during oxidation vs. electrode potential from the electrolysis of -3 -3 o a solution containing 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C using an Au/C 10 wt. % electrode...... 149 - 2 5. 6 Cyclic voltammogram for BH4 oxidation at a Au/C RDE (0.5 cm ) in the absence of surfactants at 23 oC...... 151

5. 7 Rate of gas generation vs. potential applied to an Au/C electrode. The solutions -3 -3 o contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C. a) without surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % SDS. ………………………………………………………………………...153 5. 8 Polarization curves from experiments applying a constant potential at the Au/C 2 -3 -3 cloth electrode (3 cm ). Solutions contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 oC: a) without surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % SDS...... 155 5. 9 Hydrogen generation rate vs. current density during electrolysis of solutions -3 -3 o containing 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C: a) without surfactants, b) with 0.001 wt. % Triton X-100, c) with 0.1 wt. % Triton X-100.157 5. 10 Different states of the surfactant molecules depending on the concentration. ... 158 5. 11 Atomic structure of the Triton X-100 molecule next to a borohydride molecule on Au(111): a) vertically adsorbed side view, b) vertically adsorbed front view, c) horizontally adsorbed side view, d) horizontally adsorbed front view. The yellow atoms correspond to Au, the grey atoms correspond to C, the pink atoms are B, the red are O atoms and the white are H atoms...... 162 5. 12 Comparison of the free energy of adsorption of borohydride ions on Au(111) in the absence and the presence of the two preferred configurations for the adsorbed Triton X-100 molecule vertically and horizontally...... 164 5. 13 Schematic diagram of the breaking up of the surface due to the presence of Triton X-100…………… ...... 165 5. 14 Comparison of the fuel utilization at the planar Au electrode in the presence and in the absence of surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430...... 167 5. 15 Comparison of the fuel utilization at the Au/C felt electrode in the absence and the presence of surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430...... 167 5. 16 Cyclic voltammogram at a stationary 0.125 cm2 Au disk electrode using a -1 -3 potential sweep rate of 10 mV s in solutions containing 0.02 mol dm NaBH4 + 3 mol dm-3 NaOH at 230C. a) no surfactant, b) with 0.1 wt. % SDS, c) 0.1 wt. % Triton X-100 and d) 0.1 wt. % FC4430...... 170

xii

5.17 Reverse scan of the cyclic voltammogram at a stationary 0.125 cm2 Au disk electrode using a potential sweep rate of 10 mV s-1 in solutions containing 0.02 -3 -3 0 mol dm NaBH4 + 3 mol dm NaOH at 23 C. a) no surfactant, b) with 0.1 wt. % SDS, c) 0.1 wt. % Triton X-100 and d) 0.1 wt. % FC4430……………………172 -3 -3 o 5. 18 Cyclic voltammogram of 0.02 mol dm NaBH4 in 3 mol dm NaOH at 23 C at the Au RDE (0.125 cm2) rotated between 400 rpm and 3600 rpm...... 173 5. 19 Cyclic voltammogram at a 0.125 cm2 Au disk electrode at a potential sweep rate -1 -3 of 10 mV s and at 400 rpm and 2500 rpm for 0.02 mol dm NaBH4 in 3 mol dm-3 NaOH in the absence and the presence of 0.001 wt. % SDS at 23 oC...... 174 5. 20 Limiting current density vs. square root of rotation speed (Levich) plot for borohydride oxidation at a gold RDE at 10 mV s-1. In a solution of 0.02 mol dm-3 -3 NaBH4 + 3 mol dm NaOH at room temperature in the absence () and in the presence of 0.00001 wt. % SDS (), 0.0001 wt. % SDS (), 0.001 wt. % () SDS, (◊) 0.01 wt. % SDS and (*) 0.1 wt. % SDS at 23 oC...... 175 - 2 5. 21 Comparison of the CV for BH4 oxidation at an Au RDE (0.125 cm ) in the absence and in the presence of 0.1 wt. % Triton X-100 at 23 oC and different rotation rates...... 177 5. 22 Cyclic voltammogram at a 0.125 cm2 Au disk electrode using a potential sweep -1 -3 rate of 10 mV s and at different rotation rates for 0.02 mol dm NaBH4 + 3 mol dm-3 NaOH + 0.1 wt. % FC4430 at 23 oC...... 178 1/2 -3 5. 23 Comparison of the Levich plot (jL vs. ω ) in a solution of 0.02 mol dm NaBH4 + 3 mol dm-3 NaOH at 23 oC in the absence () and in the presence of 0.1 wt. % Triton X-100 () and 0.1 wt. % FC4430 (). A gold RDE (0.125 cm2) was used at a potential sweep rate of 10 mV s-1...... 179 5. 24 Reciprocal current density vs. the inverse of the square root of the rotation rate -3 -3 for a solution containing 0.001 wt. % SDS, 0.02 mol dm NaBH4 + 3 mol dm NaOH at 23 oC and : () -0.25 V vs. Hg/HgO, (*) -0.2 V vs. Hg/HgO, ()-0.15 V vs. Hg/HgO, () 0 V vs. Hg/HgO...... 184

-3 5.25 Log ka vs. Potential for a solution containing 0.001 wt. % SDS, 0.02 mol dm -3 NaBH4 in 3 mol dm NaOH…………………………………………………..187

xiii

- 2 5. 26 CV of the BH4 oxidation at the Au/C RDE (0.5 cm ) in the absence (black lines) and in the presence (orange lines) of 0.1 wt. % SDS at rotation rates from 0 (continuous lines) to 2500 rpm (broken lines) and at 23 oC...... 190 -3 - -3 o 5. 27 Cyclic voltammogram of 0.02 mol dm BH4 in 3 mol dm NaOH at 23 C at the Au/C RDE (0.5 cm2) in the absence (black lines) and in the presence (orange lines) of 0.1 wt. % Triton X-100 at rotation rates of 0 rpm (continuous lines) and 2500 rpm (broken lines)...... 191 - 2 5. 28 Cyclic voltammogram of the BH4 oxidation at the Au/C RDE (0.5 cm ) in the absence (black lines) and in the presence (red lines) of 0.001 wt. % Triton X-100 at rotation rates at 0 rpm (continuous lines) and 2500 rpm (broken lines) at 23 oC.192 - 2 5. 29 Cyclic voltammogram for BH4 oxidation at a Au/C RDE (0.5 cm ) at rotation rates of 0 rpm (continuous lines) and 2500 rpm (broken lines) in the absence (black lines) and in the presence of 0.1 wt. % FC4430 (blue lines) and 0.3 wt. % FC4430 (green lines) at 23 oC...... 193

6. 1 SEM image of a gold-coated RVC 80 ppi prepared by a sputtering technique over 3 min, the inset shows the thickness of the deposit on the carbon substrate [109].197 −3 −3 6. 2 Cyclic voltammograms of 0.02 mol dm NaBH4 in 3 mol dm NaOH on a 0.125 cm2 gold disk electrode and on a 10 ppi RVC gold-coated electrode (1 × 0.5 × 0.5 cm) at coating time of one minute at different potential sweep rates and 298 K.199 6. 3 Potential at the peak current vs. natural logarithm of the potential sweep rate for 1 minute PVD sputtering gold deposited RVC electrodes...... 201 6. 4 Peak current vs. potential sweep rate for 1 minute PVD sputtered gold coated RVC electrodes...... 203 6. 5 Peak current vs. potential sweep rate for 2 minutes PVD sputtering gold deposited RVC electrodes...... 203 6. 6 Peak current vs. potential sweep rate for 3 minutes PVD sputtering gold deposited RVC electrodes...... 204

-3 -3 7. 1 Comparison of the hydrogen generation from 4 mol dm NaBH4 in 0.5 mol dm NaOH using a Pd/Ir catalyst using:  tap water,  river water and ■ distilled water to prepare the solution...... 213

xiv

-3 -3 7. 2 H2 generation from 4 mol dm (15 wt. %) NaBH4 in 0.5 mol dm (2 wt. %) NaOH using nanoparticulate platinum supported on Vulcan carbon printed on Toray carbon paper. The solution was prepared with distilled water. Initial temperature: 23 oC (296 K). Final temperature: 32 oC (305 K). No measurements were taken during the experiment on this occasion...... 215 3 7. 3 H2 flow rate when 10 g of Pd/C catalyst were used with 350 cm of solutions -3 -3 containing 4 mol dm NaBH4 in 0.5 mol dm NaOH prepared with:  still drinking water, and × river water. And the 350 cm3 of solution containing 4 mol -3 dm NaBH4 and no NaOH prepared with:  distilled water...... 217

7. 4 Pressure measured during H2 generation when 10 g of Pd/C were used with 3 -3 -3 solutions of 350 cm containing 4 mol dm NaBH4 in 0.5 mol dm NaOH and prepared with:  still drinking water and × river water. And the 350 cm3 of -3 solution containing 4 mol dm NaBH4 and no NaOH prepared with:  distilled water…………… ...... 218 3 7. 5 Measurement of the H2 flow rate from 350 cm of a solution containing 4 mol -3 dm NaBH4 and without NaOH in distilled water and different weights of catalyst.  10 g Pd/C  15 g Pd/C and  20 g Pd/C...... 220

7. 6 Pressure measured during H2 generation when different catalyst loading:  10 g Pd/C  15 g Pd/C and  20 g Pd/C and 350 cm3 of a solution containing 4 mol -3 dm NaBH4 without NaOH prepared with distilled were used...... 221 3 7. 7 Flow rate of H2 and pressure when 15 g Pd/C catalyst were used with 350 cm of -3 a solution containing 4 mol dm NaBH4 in the absence of NaOH (, ) and in the presence of 0.5 mol dm-3 NaOH (, ) all prepared with distilled water. The dark grey arrows show the time at with the solution was added into the upper part of the reactor...... 224 st 7. 8 Comparison of the flow rate of H2 when 20 g Pd/C was used three times:  1 use,  2nd use and  3rd use. In all the experiments 350 cm3 of deionised water -3 without stabiliser was used to prepare 4 mol dm NaBH4...... 225 st nd rd 7. 9 H2 pressure when 20 g Pd/C was used three times:  1 use,  2 use and  3 use. In all the experiments 350 cm3 of deionised water without stabiliser was -3 used to prepare 4 mol dm (15 wt. %) NaBH4...... 226

7. 10 Flow rate of H2  and pressure  when a mixture of sodium borohydride and Pd/C catalyst were mixed together at the upper part of the reactor and still drinking water was gradually added...... 228 7. 11 SEM images of: a) the unused Pd/C. The inset shows a higher magnification of the centre of the main image. b) the palladium particles deposited on one of the pore wall of the unused granulated carbon catalyst.c) the three times used Pd/C catalyst. The inset shows an enhancement image from the centre of the main picture. d) the locations of the EDX analysis of the Pd/C catalyst which had been used three times...... 231 7. 12 Nitrogen adsorption isotherm of the surface of the Pd/C catalyst at 77 K  adsorption and  desorption. Inset: Linear transform of adsorption isotherm for determination of BET surface area...... 233 7. 13 BJH pore size distributions of Pd/C catalyst determined from both:  adsorption

and  desorption branches in the N2 adsorption isotherm...... 235

xvi

DECLARATION OF AUTHORSHIP

I, Irene Merino Jimenez, declare that the thesis entitled

Investigation of the use of sodium borohydride for fuel cells and the work presented in the thesis are both my own, and have been generated by me as the result of my own original research. I confirm that:

 this work was done wholly or mainly while in candidature for a research degree at this University;  where any part of this thesis has previously been submitted for a degree or any other qualification at this University or any other institution, this has been clearly stated;  where I have consulted the published work of others, this is always clearly attributed;  where I have quoted from the work of others, the source is always given. With the exception of such quotations, this thesis is entirely my own work;  I have acknowledged all main sources of help;  where the thesis is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself; parts of this work have been published as:

1. Improvements in direct borohydride fuel cells using three-dimensional electrodes.

Ponce de León, C. Kulak, A. Williams, S. Merino-Jiménez, I. Walsh, F. C. Catalysis

Today Volume 170, 2011, Pages 148-154.

2. Developments in Direct Borohydride Fuel Cells and Remaining Challenges.

Merino-Jiménez, I. Ponce de León, C. Shah, A. F.C. Walsh. Journal of Power

Sources Volume 219, 1 December 2012, Pages 339-357.

xvii

3. Chapter 20. Developments in electrodes, membranes and electrolytes for the

direct borohydride fuel cell (DBFC). Merino-Jimenez, I., Ponce de Leon, C., Walsh,

F. C. Accepted for publication in the book: Advance materials for clean energy.

4. The design and characterization of a standalone hydrogen generator operating

via catalytic borohydride decomposition. Merino-Jimenez, I., Ponce de Leon, C.,

Walsh, F. C. Submitted to the International Journal of Hydrogen Energy.

5. The use of surfactants to suppress hydrolysis of borohydride in direct

borohydride fuel cells. Merino-Jimenez I., Ponce de Leon, C., Walsh, F. C. Ready to

submit to the Journal of Power Sources

6. The effect of surfactants on the kinetics of borohydride oxidation and hydrolysis

in the DBFC. Merino-Jimenez, I., Ponce de Leon, C., Walsh, F. C. Submitted to

Electrochimica Acta

7. Review paper on H2 generation from sodium borohydride. Merino-Jimenez, I.,

Ponce de Leon, C., Walsh, F. C. In preparation.

8. First principle study of the Pd-Ir alloys for borohydride oxidation and

comparison with experimental data. Merino-Jimenez, I. Janik, M.J., Ponce de

Leon, C., Walsh, F. C. Expected submission date: February 2014.

9. Electrocatalyst design for direct borohydride oxidation guided by first principles.

Rostamikia, G., Patel, R.J., Merino-Jimenez, I., Hickner, M., Janik, M.J. Expected

submission date: February 2014.

Siged: ………………………………………………………………………..

Date:…………………………………………………………………………….

xviii

Acknowledgements

Foremost, I would like to express my sincere gratitude to my supervisor Dr. Carlos Ponce de Leon Albarrán for his continuous support, patience and the expertise he has shared with me. His supervision and constant help has been invaluable on both an academic and a personal level, for which I am extremely grateful. I would like to thank my advisor Professor Frank Walsh for his guidance and help during my Ph.D and the writing of this thesis. I wish to acknowledge the financial, academic and technical support of the Faculty of Engineering and the Environment at the University of Southampton.

In addition, I would like to acknowledge Dr. Michael J. Janik and his research group at Penn State University for hosting me as a visiting student. I am deeply thankful to Dr. Janik for teaching me everything I know about DFT and for keeping his supervision and support for several months after I came back. I also wish to thank Stuart Barnes, Ricarda Nothelle and Saverio Gellini from Prometheus Wireless Limited for giving me the opportunity of being involved in such an interesting and challenging project, and for the financial support provided for the building and testing of the hydrogen generator.

I wish to thank Professor Pilar Herrasti from Universidad Autonoma de Madrid for her advice and support, and for indirectly bringing me the opportunity to start this Ph.D. I am very grateful to Dr. Julian Wharton for his proof-reading and useful comments during my Nine-month and Transfer Report assessments. In addition, a thank you to Dr. Richard Wills, Dr. John Low and my laboratory and office collegues Rachel, Mihaela, Ruben, Recep, Natalya, Maria, Derek, Badr and all the others I am not naming here, for their help, frienship and support.

Last, but by no means least, I would like to thank my boyfriend Giorgos for his personal support and great patience at all times. My parents, brothers and sisters-in-law have given me their unequivocal support throughout, as always, for which my mere expression of thanks likewise does not suffice. I thank my friends in United Kingdom,

xix

Spain and elsewhere for their unconditional support and encouragement, which have meant a lot to me over the past three years.

Abbreviations

BE: Binding energy

BET: Brunauer, Emmett and Teller

BJH: Barrett-Joyner-Halenda

CCS: Colloidal carbon spheres

CHMC: Critical Hemi-Micellar Concentration

CI-NEB: Climbing image nudged elastic band method

CMC: Critical micelle concentration

CMetC: polymer- carboxymethyl cellulose

CTAB: Cetyltrimethylammonium bromide

CV: Cyclic voltammetry

DBFC: Direct borohydride fuel cell

DFT: Density functional theory

EC: Electrochemical chemical

ECE: Electrochemical chemical electrochemical

ETFE-g-PSSA: polyethylenetetrafluoroethylene

FC: Fuel cell

FeTMPP: tetramethoxyphenyl prophyrin fcc: Face centred cubic

GC: Glassy carbon

GGA: generalized gradient approximation

HTES: high-temperature electrolysis of steam

IBFC: Indirect borohydride fuel cell

MEA: Membrane electrode assembly

xxi

ML: Monolayer

OCV: Open circuit voltage

PEM: Proton exchange membrane ppi: linear pores per inch

PVD: Physical vapour deposition

RDE: Rotating disc electrode

RVC: Reticulated vitreous carbon

SCE: Saturated calomel electrode

SDS: Sodium dodecyl sulphate

SEM: Scanning electron microscope

SHE: Standard hydrogen electrode

TEAH: Tetraethylamonium hydroxide

TU: Thiourea

VASP: Vienna ab initio simulation program

Wt. %: weight percent

ZPVE: Zero-point vibrational energy

Symbols

A Geometrical surface area of the electrode cm2

-3 c Bulk concentration mol cm

2 -1 D Diffusion coefficient cm s

E Electrode potential V

Ecell Cell voltage

o E cell Open-circuit cell voltage

xxii

EP Oxidation peak Potential V f constant equal to F/RT = 38.92 V-1

-1 F Faraday constant C mol

G Gibbs free energy eV h Plank constant eV s

̂ Hamiltonian operator

I Current A

j Current density A cm-2

-2 jL Limiting current A cm

-1 ka Heterogeneous oxidation rate constant cm s

-1 Ks Standard rate constant for the electron-transfer cm s m Particle mass g n Number of electrons exchanged in the reaction

( ⃗) Electronic density

N Number of electrons in a molecule

P Power density W cm-2 r Spatial coordinate

Re Electrical resistance W

R Gas constant J mol-1 K-1

S Entropy J K-1

T Temperature K

U Absolute (vacuum reference) electrode potential V

Vo Equilibrium potential V

Vs External potential V

xxiii

External potential for a virtual non-interacting system V

Exchange-correlation potential V

Hartee potential of the electrons V

 Charge transfer coefficient

Symmetry factor

Kohn-Sham eigenvalue eV

hact Activation overpotential V

hconc Concentration overpotential

μ Kinematic viscosity cm s-1

ν Potential sweep rate V s-1

Φ Energy at the ground state eV

Hartree energy eV

({ }) Coulombic energy associated with interactions among the nuclei (or ions) at positions { } eV

Kinetic energy eV

Potential energy eV

Exchange-correlation energy eV

Wave function

ω Rotation rate rad s-1

Vector of partial derivatives (∂f / ∂x1, ∂f / ∂x2…, ∂f / ∂xn)

xxiv

Chapter 1: Introduction

1.1. Background and motivation

The global concern with climate change and the environmental impact of the consumption of fossil fuels has become increasingly apparent during the past four decades. As a consequence, the research for new renewable, clean, safe and efficient sources of energy has substantially been investigated. Fuel cells offer a promising alternative to incumbent electrical power generation technologies (primarily based on fossil fuels), for large-scale applications (i.e. remote or backup power), as well as small- scale applications (i.e. laptops or mobile phones). Hydrogen is suitable as a clean energy source for electronic devices due to its highly rich energy per unit weight, 120 MJ kg-1, which is three times higher than that of gasoline, 44 MJ kg-1 [3]. However, the most highly developed fuel cell, namely the direct hydrogen-oxygen fuel cell, presents problems related to the sourcing, storage and safe handling of hydrogen (particularly for mobile applications) [4].

The problems are due to the hydrogen properties: it is a highly volatile and flammable gas, with limits between 4 % and 75 % vol. hydrogen. It is potentially explosive in air; the explosion limits are related to the temperature and pressure of the system, being non explosive at any pressure for temperatures below 400oC (673 K) [5]. However, it has extremely low ignition energy (0.017 mJ for mixtures with air and 0.012 mJ for mixtures with oxygen [6]), a low viscosity, and high combustion and detonation velocities. For these reasons, hydrogen must be treated and stored very carefully [7].

Besides the safety issues, the poor energy per volume basis of the hydrogen gas implies the need of large volumes to be able to store the amount of hydrogen required to feed a

1 Chapter 1: Introduction

fuel cell [8], i.e. a tank of 225 dm3 stores 4 kg of hydrogen gas at 200 bar pressure. [7,

9]. Despite this low volumetric energy density, researches in hydrogen storage have been developed in the past years with new hydrogen storage techniques [10]. Nowadays hydrogen can be stored in three forms, hydrogen gas in steel cylinders at a high pressure of 150 bar - 200 bar and at room temperature is the most common storage system. New lightweight composite cylinders able to withstand pressures up to 80 MPa (800 bar), have been developed, reaching a volumetric hydrogen density of 36 kg·m−3, which is approximately half of that in its liquid state. Hydrogen can also be stored as a cryogenic liquid, at 21.2 K (-252 oC) and ambient pressure. The third hydrogen storage system is by physisorbed /chemisorbed in solids, which among the hydrogen storage forms currently available is probably the most suitable for portable applications. Many metals and alloys are capable of reversibly absorbing large amounts of hydrogen, but the highest hydrogen volumetric densities are obtained from metal hydrides [10-12], such as

LiH, LiAlH4, LaBH4, NaH, NaBH4, which can store large amounts of hydrogen in a non-reversible way. These metal hydrides react with water to generate hydrogen through their hydrolysis and a catalyst can be used to control their reaction rate [13].

The hydrogen generated is then oxidised in the anode:

(1.1)

Among the metal hydrides, the borohydrides have very high storage density capability due to its large hydrogen content. Figure 1.1 shows the hydrogen content in wt. % of several hydrides, and it can be appreciated that lithium borohydride, LiBH4, has the highest hydrogen content 18.5 wt. % followed by sodium borohydride 10.6 wt. %.

Lithium borohydride has a high thermal stability temperature (~ 400 oC) and hydrogen

2 Chapter 1: Introduction

is only released under extreme pressures (70 bar -350 bar) and temperatures (600 oC -

650 oC) [3].

6 Hydrogen conent,Hydrogen % wt.

0 LiBH4 NaBH4 LiAlH4 MgH2 NaAlH4 CaH2 LiBH4 NaBH4 LiAlH4 MgH2 NaAlH4 CaH2

Figure 1. 1 Diagram of the comparison of the hydrogen content (wt. %) of different hydrides.

Besides sodium borohydride contains less hydrogen, its stability and easy handling, as well as the possibility of operation at ambient conditions, makes it preferable for hydrogen generation rather than lithium borohydride [14]. Borohydride hydrolysis is an exothermic reaction that liberates 267 kJ mol-1 and 90 % of the stoichiometric amount of hydrogen predicted from the hydrolysis reaction (1.2) [14] [15]:

( ) (1.2)

Where x is the hydration factor, which could be > 2 forming the hydrated of the borate,

NaB(OH)4 or < 2 leaving solid by-products and unreacted borohydride [16]. The

3 Chapter 1: Introduction

hydrogen can be fed into a H2/O2 fuel cell: reaction (1.2) indicates that if the borohydride is completely consumed (x > 2) four moles of hydrogen could be released per mole of borohydride consumed, i.e. 89 litres of H2 at 25 ºC (298 K) temperature and

1 atmosphere pressure, if reaction (1.2) is 100 % efficient. In practice 100% efficiency is rarely achieved, in part because the by-product causes an increase in the pH of the solution [14, 15]. Figure 1.2 shows the indirect borohydride fuel cell system (IBFC), where the sodium borohydride reacts with water in a separate catalytic reactor and releases hydrogen, which will feed a H2/O2 fuel cell. The IBFC has been commercialised by one company with a guaranteed operation time of 500 hours; a continuous output power of 200 W at 10 A current; a stack cell voltage of 20 V - 32 V; and a total specific energy of 450 W h kg–1 – 607 W h kg–1 [17].

H O Pump Heat 2 exchanger H2

Silica O2 Hydrogen generator drier containing a catalyst Fuel Cell

Gas- H2 gas NaBH4 liquid solution tank separator

Borate Solutions

Figure 1. 2 Arrangement of an indirect borohydride fuel cell (IBFC) system composed by a hydrogen generator from NaBH4, which product feeds a H2/O2 fuel cell [1].

4 Chapter 1: Introduction

Another alternative solution to the aforementioned issues related to hydrogen safety, storage and transportation, is the replacement of the gaseous hydrogen in the fuel cell with a liquid hydrogen carrier, such as ethanol or methanol, which can be directly oxidised in the anode, through reactions (1.3) and (1.4):

o E = 0.74 V vs. SHE (1.3)

o E = 0.81 V vs. SHE (1.4)

The latter leads to the direct methanol fuel cell (DMFC), which suffers from high rates of reactant crossover and low power densities. Alternatively, a solution of one of the metal hydrates, such as NaBH4, can be directly oxidised in the direct borohydride fuel cell (DBFC) [18]:

E = -1.24 V vs. SHE (1.5)

The use of sodium borohydride (NaBH4) as a fuel for fuel cells presents many advantages: it is stable in strong alkaline solutions with a half-life of approximately 430 days at pH 14; is available as a solid or as a 30 wt. % solution [4]; it contains a large mass concentration of hydrogen (10.6 wt. %) stored in a safe and innocuous form [19]; it is theoretically able to release 8e– per borohydride ion at a very low electrode potential (–1.24 V vs. SHE) and is able to generate a high power output using a small quantity of fuel [20]; it does not produce carbon dioxide when it is used in fuel cells, unlike ethanol or methanol [21], or when used for hydrogen production, unlike hydrocarbons such as natural gas; its oxidation product, metaborate, is environmentally acceptable and can potentially be recycled to sodium borohydride [22]; its availability to

5 Chapter 1: Introduction

operate at ambient conditions and the safety of the reactants and products makes it suitable for portable applications.

Table 1.1 shows the theoretical energy densities for different fuels and oxidants calculated at 25 oC and 1 atm. Comparing these values, the maximum energy

-1 corresponds to the NaBH4/H2O2 system, 17 kW h kg , followed by the NaBH4/O2, 9.3

-1 kW h kg . The NaBH4/H2O2 theoretical energy density is more than five times higher than that for H2/O2 fuel cell and two and three times higher than those of C2H5OH/O2 and CH3OH/O2, respectively. It should be considered that these are theoretical values and in practice these values are currently not achievable due to the limitations in concentration, thermodynamic and kinetic aspects, which will be further discussed in the next chapter.

Theoretical energy output Fuel Oxidant Reference / kW h kg-1

- BH4 H2O2 17 [23] - BH4 O2 9.3 [24]

Ethanol (C2H5OH) O2 8.04 [25]

Methanol (CH3OH) O2 6.08 [24]

H2 O2 3.28 [26]

Table 1. 1 Theoretical energies of various fuels and oxidants, assuming complete oxidation of the fuel.

Figure 1.3 shows the direct borohydride fuel cell (DBFC), where the sodium borohydride is directly oxidised at the anode. The best performance documented for a

DBFC is a power output of 36 W obtained by Miley et al. using a 25 cm2 (1.44 W cm–2)

6 Chapter 1: Introduction

NaBH4/H2O2 single cell with a Pd/C anode and Au cathode separated by a Nafion 112 membrane [18]. However, further research is required to commercialize it due to the low efficiency obtained in practice compared with the theoretical. The following sections describe the most important electrocatalysts, reactor and cell designs used in both the DBFC and the IBFC.

Anode Cathode Anolyte Reservoir 1M - + Catholyte Cell 1M Reservoir NaBH 4 H2O2

1M 3M HCl NaOH

Reference electrode

Flow meters

Figure 1. 3 Arrangement of a direct borohydride-hydrogen peroxide fuel flow cell system (DBFC) [23].

7 Chapter 1: Introduction

1.2. Aims and objectives

The aim of the thesis is to increase the understanding of the borohydride oxidation and its hydrolysis, and to investigate the effect of different catalysts, the presence of surfactants and the operation conditions in both oxidation and hydrolysis reactions. The research therefore pursed the following objectives:

 Evaluation of different anodic electrodes for the borohydride oxidation including

planar Au, Au/C, gold coated reticulated vitreous carbon (RVC) and Pd-Ir alloy,

in order to find an appropriate catalytic material towards the borohydride

oxidation and non-catalytic towards the borohydride hydrolysis.

 Analysis of the effect of surfactants in the borohydride hydrolysis and oxidation

of borohydride ions, including their effect in the kinetics of the borohydride

oxidation and its hydrolysis. A quantitative analysis of the hydrogen generated at

constant potentials in the presence and absence of surfactants is performed.

 Design, construction and testing of a reactor to generate hydrogen from sodium

borohydride. Evaluation of different catalyst for the hydrogen generation from

the borohydride hydrolysis reaction, such as platinized titanium, Pd-Ir alloy, Pt

nanoparticles on carbon paper and Pd deposited on granular carbon.

8 Chapter 1: Introduction

1.3. Thesis outline

This thesis is divided into 8 chapters, including the introduction and literature review

(Chapters 1 and 2), the experimental description (Chapter 3), the analysis and discussion of the results (Chapters 4 to 7) and the conclusions and suggestion for future work

(Chapter 8). In Chapter 2, a thorough literature review of the use of sodium borohydride for fuel cells is presented, including a deep study of both the oxidation reaction and the borohydride hydrolysis, for the direct and indirect borohydride fuel cells applications.

Chapter 3 describes the experiments designed to address the thesis aims, mentioned in the previous section of this chapter. Chapter 4 includes the results and analysis of the experimental work on the use of a Pd-Ir alloy as an anode material for the borohydride oxidation, together with the mathematical studies to analyse the mechanism of reaction of the borohydride oxidation at the Pd-Ir electrode. Density functional theory (DFT) studies were carried out in collaboration with Dr. Michael Janik at Penn State

University. This chapter also includes the comparison of the free energy of adsorption of hydrogen on Pd(111), Pd2Ir1 and Pd2Ir2. The structure of Pd(111) was built and optimized by Tomas Senftle from Penn State University, and that structure was used for the calculations of the free energy of adsorption of hydrogen at low and high coverage.

In Chapter 5, the results from experimental work regarding the effect of the use of surfactants in the borohydride hydrolysis and oxidation reactions at gold electrodes are presented. DFT studies on the free energy of adsorption of the borohydride ions at the gold electrode and the effect of the presence of the adsorbed surfactant Triton X-100 was analysed. For the optimization of the structure of the gold material, a cell of 3 × 6 ×

1 was built and optimized. To build such a cell, a smaller cell consisting of 3 × 3 × 1, 9 Chapter 1: Introduction

which was previously built and optimized by Dr. Reza Rostamikia from Penn State

University, was doubled in size and optimized again. Chapter 6 includes the analysis of the borohydride oxidation at three-dimensional (reticulated vitreous carbon) gold coated electrodes by physical vapour deposition.

In Chapter 7, the experimental work is focused on the use of sodium borohydride to generate hydrogen gas. A hydrogen generator from sodium borohydride is built and different catalysts such as Pd-Ir alloy, Pt nanoparticles on carbon paper and Pd deposited on granular carbon (Pd/C) are tested with the objective to generate enough hydrogen gas to feed a H2/O2 fuel cell. The experiments about nitrogen adsorption on the Pd/C catalyst, which results are described in section 7.5, were performed by Ruben

Porras. Finally the results are drawn together into a set of conclusions in Chapter 8, and suggestions for further work are provided.

10 Chapter 2: Literature Review

Chapter 2: Literature Review

Figure 2.1 shows a histogram of the number of papers published on the borohydride oxidation/DBFC and the borohydride hydrolysis. The first studies on borohydride oxidation for fuel cells started in 1960, when Pt and Ni were investigated as anode catalysts by Elder et al. [27] and Indig et al. [28], respectively. An increasing interest in the field over the last ten years is clearly apparent from this plot. The number of publications on direct borohydride fuel cells maintained an average of 53 papers per year between 2007 and 2012 which corroborates the interest in this technology. Much of the focus has been placed on the hydrolysis of borohydride ions [29], due to its relevance to both direct and indirect borohydride fuel cells. Indeed, the number of publications on borohydride hydrolysis is slightly higher than those on the DBFC itself, mainly in the past two years.

70 DBFC 70 60 borohydride 60 hydrolysis 50

40 30

30 10

Number Number publicationsof 20

1963 1955 1957 1959 1961 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 1953 Year

Figure 2. 1 Evolution of the number of publications related to the direct borohydride fuel cell (DBFC) and the hydrolysis of borohydride. The inset shows a close up between the years 2001 and 2012 [2]. 11

Chapter 2: Literature Review

2.1. Sodium borohydride as a fuel: Direct Borohydride Fuel Cell

2.1.1. Anode

This section describes the problem of finding an appropriate anode material for the direct borohydride fuel cell and summarises the main catalysts reported in the literature.

- The choice of anode catalyst for BH4 oxidation has been widely investigated.

Platinum, nickel, and palladium are known to be more active for borohydride oxidation reactions; however an extreme activity of these catalysts for breaking B-H bonds leads to large amounts of hydrogen gas production and less than 8e- collected per borohydride molecule converted. This lack of selectivity also limits the overall efficiency of the fuel cell. Materials such as Au or Ag are considered the best options for the DBFC.

However, due to the high cost involved, investigations have also been carried out in the use of metal alloys instead of pure metals.

At the anode of the DBFC, the complete oxidation of borohydride ions to release 8e– takes place, according to reaction (1.5) [27]. The actual number of electrons released is typically less than the theoretical 8, due to the parallel reaction of hydrolysis, which competes with the oxidation and depletes the borohydride ions available:

(2.1)

This reaction is favourable when the electrolyte is not a strong alkaline solution, and in the presence of certain catalysts such as Pt, Ru, Ni, NixB, Co, CoxB or Pt-LiCoO2 [30].

Hence, considering that reactions (1.5) and (2.1) take place at the same time, the actual

12 Chapter 2: Literature Review

electrochemical reaction occurring at the electrode surface can be represented as follows:

( ) ( ( ⁄ ) ) (2.2)

where x is the number of electrons released from each borohydride ion and is also the number of hydroxide ions involved [31]. When x = 8, only the borohydride oxidation will take place and reaction (2.2) will be equal to reaction (1.5). When x = 0, no electron transfer takes place and reaction (2.2) can be replaced by reaction (2.1). The

– competition between the BH4 direct oxidation reaction and hydrolysis is a function of the electrode material, the electrolyte composition and the electrode potential [2]. It has

– – been demonstrated that for a ratio of OH to BH4 concentrations of approximately 4.4, the oxidation releases 8 electrons, while for lower ratios the concentration of protons increases, lowering the pH and increasing the concentration of intermediate species

– (BH3OH ), which leads to a lower number of electrons been liberated [32]. The precise mechanism of borohydride oxidation is still subject to speculation and, therefore, awaits more detailed investigation. The following description represents the current understanding of the oxidation steps. The initial step is an electron transfer followed by a rapid decomposition of a radical and a further electron transfer step; i.e. an ECE

– sequence. This oxidation mechanism of BH4 is observed in metals with non-catalytic surfaces for hydrogen adsorption, such as Au or Ag, involving an initial electron

· transfer to form an unstable BH4 radical. This first step of the mechanism was suggested on the basis of cyclic voltammetry experiments, at 200 V s–1 on Au microelectrodes [33]:

13 Chapter 2: Literature Review

(E) (2.3)

(E) (2.5)

The second stage, which is thought to involve 6e– in a faster electrochemical reaction, could not be experimentally determined. The only investigation which attempted to elucidate the second stage of the borohydride oxidation was carried out by Mirkin et al.

[33] who performed computational studies to simulate the cyclic voltammetry of the borohydride oxidation at an Au surface at 104 V s-1. For scan rates at or above 500 V s-1, significant deviations between experiment and theory occurred. Thus, the remaining steps of the borohydride oxidation are still subject to study. The reaction mechanism may change depending on the choice of catalyst and reaction conditions, such as the concentration of borohydride or the pH [31].

The mechanism of reaction (2.2) is likely to involve an initial pre-disassociation step at the active surface sites, for example, at a platinum electrode. The borohydride ions react on the electrode surface liberating one proton [34, 35]:

(2.6)

This step could be followed by the adsorption of hydrogen ions on the metallic surface, particularly in metals M, that are able to promote the pre-dissociation step, such as Pt,

Pd or Ni [4]:

(2.7)

14 Chapter 2: Literature Review

This reaction is followed by an electron transfer step (E), a further chemical surface reaction (C) and the combination of hydrogen atoms producing hydrogen gas [31]. The oxidation of borohydride ions at Pt electrodes is complicated by the catalytic hydrolysis and oxidation of the intermediates that occur simultaneously. The CE mechanism of

– BH4 at Pt follows the sequence below [20]:

(E) (2.9)

(E) (2.10)

– The intermediate product of the borohydride oxidation (i.e. BH3OH ions) can be either oxidized or hydrolyzed, and the number of electrons obtained from one molecule of borohydride can vary depending on the route of the reaction mechanism. The electron-

– transfer reaction rate during the formation of borate ion (BO2 ) (reaction 2.10), varies with the nature of the electrode material depending on whether or not it catalyzes the hydrogen evolution reaction (2.8).

– - In most studies, the BH3OH ion has been proposed as an intermediate during BH4 oxidation at Pt, Pd, Au, or Hg anodes. This intermediate was considered to be bound to

Pt or Au surfaces [36]. Gyenge [32] has proposed that the oxidation pathways may

– depend on the surface adsorption of both BH4 and various reaction intermediates. The irreversibility of the borohydride oxidation is the result of very unstable intermediates such as the short-lived radical BH4˙ [27]. Figure 2.2 shows a sketch of the different pathways of the mechanism of oxidation of the borohydride ions depending on the materials and alkalinity of the borohydride solution [35, 37].

15 Chapter 2. Literature Review

Direct anodic oxidation 6 < x < 8.

Non-catalytic ( ) ( ( ⁄ ) ) Mechanism: surfaces for H2 adsorption e.g. Hg, st 1 Step Ag, Au Electron transfer

0 < x < 3

nd 2 Step Fast electrochemical reaction, still under investigation (Phase

homogeneous)

Catalytic surfaces for ⁄ H2 adsorption e.g. Pt, Pd, Ni, Co, Ru Electron transfer 2 < x < 6

⁄ ( ) ( ) ⁄

⁄ ( )

( )

( ) ( )

⁄

Figure 2. 2 Reaction pathways involving anodic oxidation of borohydride ion and competitive reactions [35, 37].

16 Chapter 2. Literature Review

Studies have been conducted on a wide range of anode catalysts in order to assess their effect on borohydride oxidation and cell performance [18, 31, 32, 38-42]. Electrode materials that are active towards borohydride hydrolysis include: Pt, Ru, Ni, NixB, Co,

CoxB, Pt-LiCoO2, whereas non-active materials for hydrolysis are Hg, Au, and Ag [30].

Non-precious materials such as Ni and Cu have also been investigated as catalysts for the anodic reaction in DBFCs [31]. Few studies have taken into account the cathode catalyst and its influence on fuel cell performance [36, 41, 43-45], which is surprising considering that the cathode reaction, usually oxygen reduction is often found to be the limiting step [46-48]. The most of the information regarding the cathode materials for

- the BH4 /O2 fuel cell has been taken from publications about the H2/O2 fuel cell, due to the similarity of the system operation in the cathodic compartment.

– Figure 2.3 shows a comparison of the power density obtained from a BH4 /O2 DBFC system of 4 cm2 active geometric area with different anode materials. In all cases, the cathode was Pt/C (2 mg Pt cm–2), a Nafion 117 membrane was used to separate the anode and cathode compartments, and the operating temperature was 85 oC with an

–3 –3 electrolyte containing 5 wt. % (1.32 mol dm ) NaBH4 in 10 wt. % (2.5 mol dm )

NaOH [49, 50]. The best performance observed in Figure 2.3, was using Pt-Ni as anode, obtaining the highest peak power density, 250 mW cm-2, even when the temperature used was lower than in all the other systems. Higher peak power density was obtained using Pd/C as anode, 89.6 mW cm–2, than using Au/Ti or Au/C, 80 mW cm–2 and 72.2 mW cm–2, respectively. The lowest peak power density was 40.5 mW cm–2, obtained with a Ni/C anode. The effect of the increase in the concentration of borohydride ions on the power density of the cell was demonstrated when a Zr-Ni laves-phase alloy

(intermetallic phases with composition AB2) was used as the anode catalyst. A solution

17 Chapter 2. Literature Review

–3 –3 containing 10 wt. % (2.64 mol dm ) NaBH4 in 20 wt. % (5 mol dm ) NaOH was fed into the anode compartment [46] and the peak power density of the cell increased to around 190 mW cm–2.

250 b)

200

2 - a)

150

/ mW cm mW /

100

Power density, density, Power 50 e) d) f) h) g) 0 0 100 200 300 400 500 600 Current density, j / mA cm-2

- Figure 2. 3 Power density (P) vs. current density in a BH4 /O2 DBFC with various 2 anodes of 4 cm active area: a) Zr0.9Ti0.1Mn0.6V0.2Co0.1Ni1.1, b) Pt-Ni, c) Pd/C, d) Au/Ti, e) Au/C, f) Ag/Ti, g) Pt/C, h) Ni/C. Cathode: Pt/C (2 mg Pt cm-2). Membrane: Nafion 117. Temperature: 85oC, except of o -3 b) 60 C. Fuel: a): 10 wt. % (2.64 mol dm ) NaBH4 in 20 wt. % (5 mol -3 -3 dm ) NaOH [24] and b) to h): 5 wt. % (1.32 mol dm ) NaBH4 in 10 wt. % (2.5 mol dm-3) NaOH [38, 39].

The fundamental aspect that still needs to be addressed for DBFCs and one of the objectives of this work is to find an appropriate anode material, active towards the borohydride oxidation with little or no catalytic activity towards borohydride hydrolysis. A description and critical analysis of the several anode materials that have been tested on the DBFC is presented following.

18 Chapter 2. Literature Review

Platinum and its alloys

The calculated number of electrons transferred during the oxidation of borohydride ions at a Pt electrode lies between 2 and 8 [27, 30, 51]. Concha et al. [51] defended that the number of electrons involved in the borohydride oxidation reaction depends on the thickness of the active layer of electrocatalysts, being nearly 8 for a layer of approximately 3 μm Pt/C and n ∼ 2 for an active layer with a thickness < 1μm. The thick Pt/C layer displays enough residence time for the molecule to complete the hydrogen oxidation reaction and/or the borohydride oxidation reaction and that will increase the number of electrons.

Density functional theory (DFT) calculations suggest that these bond breaking reactions

– are slower on gold than on platinum surfaces indicating higher overpotential for BH4 oxidation at Au surfaces [52]. This seems to agree with the experimental observation that the heterogeneous rate constant obtained for the four-electron transfer during the

– direct oxidation of BH4 at Pt is ten-fold that for the eight-electron oxidation on Au

[32].

Several binary Pt alloys have been tested: Pt-Ir, Pt-Ni, Pt-Au, Pt-Ru/C and Ag-Pt [30].

Tegou et al. [53] tested a Au-Pt mixed coated on Ni-modified glassy carbon substrates by galvanic replacement of Ni electrodeposits upon immersion in a mixture of chlorolaurate and chloroplatinic acid. The alloy showed catalytic activity towards the borohydride oxidation with an apparent number of electrons of 7.1 and kinetic currents closer to those of pure Pt than pure Au. Gyenge et al. [30] synthesised and investigated a bimetallic Pt-Au catalyst that could theoretically combine the favourable kinetics on

Pt with the higher coulombic efficiency for borohydride oxidation on Au. Experiments

19 Chapter 2. Literature Review

–3 –3 with a DBFC were carried out using a 2 mol dm NaBH4 in 2 mol dm NaOH solution on the anode side, 5 mg cm–2 anode loading of Pt-Au alloy, a Nafion 117 membrane,

–2 and an O2 gas diffusion cathode containing 4 mg cm Pt. The results obtained showed that the peak current was higher, and that the oxidation peak potentials were shifted to more negative values on Pt-Au compared to pure Pt, making easier the oxidation reaction. The number of electrons transferred during the borohydride ion oxidation was

8 on the Pt-Au alloy. Pt-Ir and Pt-Ni were also tested as working electrode catalysts for the same cell. The results showed that Pt-Ir and Pt-Ni were the most active anode catalysts, giving in both cases a power density of 53 mW cm–2 at 75 oC (348 K). Pt-Ir showed potentially favourable kinetics, with the oxidation peaks shifted to more negative potentials than those observed with Pt-Ni, and yielding the highest

– voltammetric BH4 oxidation current densities at potentials more negative than –0.4 V vs. mercury oxide reference electrode, which is ultimately the domain of interest for borohydride fuel cells. Both, Pt-Ir and Pt-Ni gave the highest cell voltages at any given current density, e.g. at 100 mA cm–2 and 52 oC (325 K) the cell voltage was 0.53 V vs. mercury oxide (Hg/HgO) reference electrode with an anode catalyst loading of 5 mg cm–2 in the anode.

Duteanu et al. [38] reported experiments using a binary alloy of Pt and Ru deposited on a carbon anode together with a Pt/C cathode in a MEA. The catalyst loading was 1 mg cm–2 in both cases. The cell gave power densities of 145 mW cm–2 and 110 mW cm–2 using oxygen and air cathodes, respectively, with a cathode flow rate of 0.4 cm3 min-1 at

60 oC (333 K) and borohydride concentrations of 1 mol dm–3 in sodium hydroxide 1 mol dm–3. Concha et al. [54] suggested that the combination of Ag and Pt could improve the cell performance since Ag promotes the direct oxidation of borohydride,

20 Chapter 2. Literature Review whereas Pt helps to increase the reaction rate. The authors tested the Ag-Pt alloy as an

–3 –3 anode for borohydride oxidation using 0.001 mol dm NaBH4 in 0.1 mol dm NaOH at

25 oC (298). The borohydride oxidation reaction was faster on Pt than on Ag, the measured current density being 150 mA cm–2 for the former and 3.1× 10–1 mA cm–2 for the latter at -0.65 V vs. SHE [54]. Two different compositions of Pt-Ag alloy were prepared, Pt-Ag composed of 94.5 % Pt and AgPt with 8.4 % Pt. The number of electrons transferred on both alloys at the limiting current was 4 and it was concluded that Ag in PtAg favours the direct pathway to the oxidation of borohydride ions, while

Pt in AgPt enhances the borohydride oxidation reaction kinetics. As a result, both Pt-Ag alloys show a unique and comparable behaviour regarding borohydride oxidation:

– unlike Pt alone; they do not catalyse the quantitative heterogeneous hydrolysis of BH4 followed by the hydrogen oxidation reaction but drive the borohydride oxidation reaction towards the direct pathway.

Gold

A number of investigations have been carried out with different stationary and rotating electrodes based on gold, and good performance for borohydride oxidation has been observed [20, 39]. The experiments demonstrate that 8 electrons are released during the direct oxidation, which means that hydrolysis does not seem to occur during the operation [33, 41, 55]. However, recent findings contradict this effect. Chatenet et al.

[56] claimed that the amount of hydrogen released at an Au electrode is not negligible and they proposed that the borohydride oxidation pathway can be different at low (E <

0.3 – 0.5 V vs. SHE) and high potential values (E > 0.3 – 0.5 V vs. SHE). They pointed out that if the hydrolysis of borohydride ions proceeds in two steps (reactions (2.8) and

(2.11)) and the first step of the mechanism of reaction of borohydride oxidation is as

21 Chapter 2. Literature Review

– shown in reaction (2.12) [33], then BH3OH can be produced through either reaction

(2.8) or reaction (2.12).

(2.11)

(2.12)

– Chatenet et al. [56] suggested that, at low overpotentials, the oxidation of BH3OH on gold can involve three electrons (reaction (2.8) followed by reaction (2.13)) to six electrons (reaction (2.8) followed by (2.14)):

(2.13)

(2.14)

− At high overpotentials, the activity of Au for the oxidation of BH3OH is important and the direct oxidation of borohydride, reaction (2.12), occurs, followed by the oxidation of

– the intermediate species (BH3OH ), reaction (2.13) or (2.14), releasing between five and eight electrons in total. It is also possible that an overall EC mechanism takes place,

– where BH3OH is formed through reaction (2.12) then hydrolysed through reaction

(2.11), in which case, more hydrogen will be generated. Finkelstein et al. [20] suggested that Pt can outperform Au for DBFCs, providing similar charge efficiency at less positive anode potentials. However, the hydrolysis problem on the Pt electrode was not considered by the author.

22 Chapter 2. Literature Review

Gold alloys such as Au-Cu, Au-Ni and Au-Co have also been considered for borohydride oxidation showing higher catalytic activity than monometallic Au catalyst

[57-59]. Yi et al. [58] tested a Au-Cu bimetallic nanoparticles supported on carbon black XC-72R (Au–Cu/C). Different proportions of Au and Cu were tested (Au/C,

Au75Cu25/C, Au67Cu33/C and Au50Cu50/C) under the same conditions, obtaining that the

− Au67Cu33/C catalyst presented the highest catalytic activity for BH4 electrooxidation showing a maximum power density of 51.8 mW cm−2 at 69.5 mA cm−2 and 20 °C (293

−2 K) compared to 19.9 mW cm when Au/C was used. Au45Co55/C as anode electrocatalyst showed higher power density, 66.5 mW cm−2, at a discharge current density of 85 mA cm−2 at 25 °C (298 K) [59]. Gold-copper alloys showed increased activity compared to a pure gold electrode, while maintaining the selectivity to direct oxidation and increasing the stability compared to pure Cu. The use of Au-based alloys as anode material for DBFC seems to be beneficial compared to pure Au, not only for the reduction of the material cost but also for the increase in activity.

Silver

According to Concha et al. [51] the number of electrons released at Ag varies from 2 at pH 12.6 to 6 at pH 13.9, at 25 oC. Ag, however, shows slow electrode kinetics and low power densities [40]. A comparison between Ag and Ag alloys (Ir, Pt, Au and Pd) was carried out by Atwan et al. [60], who suggested that alloying Ag with Ir and Pt could effectively improve the electrode kinetics, decrease the overpotential and obtain a more negative open circuit potential, leading to less catalytic activity towards the borohydride hydrolysis.

23 Chapter 2. Literature Review

It is also necessary to take into account the presence of oxides on the Ag surface.

Studies related to the formation of oxides on Ag electrodes in alkaline media [61, 62] show that Ag oxide layers were formed and present during borohydride oxidation [61].

Sanli et al. [62] have also shown that oxides formed on the Ag surface by cyclic voltammetry have a catalytic effect on the oxidation of borohydride. The oxidation of borohydride appears to be promoted at high pH due to the formation of a multi-layered oxide film (Ag2O) on the catalyst surface. They found that the number of electrons transferred on silver oxide was 6 and suggested the following anodic reaction:

(2.15)

Nickel

The oxidation of borohydride ions on Ni involves the transfer of 4 electrons, (the other

4 electrons are lost in the formation of hydrogen gas), although high current densities can be reached at an acceptable potential (e.g. 300 mA cm–2 at -0.7 V vs. SHE) [4].

(2.16)

Liu et al. [31] experimented with Ni, Raney Ni, Pd, Pt, Cu and Au as anode catalysts in

–3 order to compare their OCVs. They used a solution containing 6 mol dm NaBH4 in 6 mol dm–3 NaOH, and a Nafion membrane in the MEA. Under the same conditions, Ni and Raney Ni electrodes exhibited the most negative OCVs of –1.03 V vs. SHE, followed by Cu and Au with values of –1.02 V vs. SHE and –0.99 V vs. SHE, respectively, and Pt and Pd with values of –0.91 V vs. SHE [31]. Higher efficiency was

24 Chapter 2. Literature Review achieved, however, on Pd and Pt electrodes under certain conditions, such as a relatively low borohydride concentration and/or a large anodic current [30, 40]. Liu et al. [63] reported the polarization curves in a cell with Ni anode catalyst, a metal hydride as the cathode catalyst, and a Nafion membrane to separate the two electrodes. The results showed values of 200 mA cm–2 at –0.7 V vs. Hg/HgO using a solution of 1.6 wt.

–3 –3 o % KBH4 (0.4 mol dm ) in 6 mol dm KOH at 20 C. Suda et al. [46] reported a cell voltage of 0.6 V at 150 mA cm–2, operating at 60 oC (333 K), when they used a Zr-Ni laves phase alloy as an anode catalyst at a loading of 200 mg cm–2. Ma et al. [64] investigated Ni composite anodes, such as Ni + Pt/C and Ni + Pd/C (the ratio of Pt-Ni or Pd-Ni was 25:1). They assembled a borohydride-oxygen fuel cell consisting of 1 mg cm–2 Pt/C cathode separated from the anode by a Nafion membrane. The anolyte was 5

–3 –3 wt. % NaBH4 (1.32 mol dm ) in 10 wt. % (2.5 mol dm ) NaOH aqueous solution, flowing at a rate of 5 cm3 min–1. The cell was operated for 100 hours to monitor its performance and stability. When carbon-supported palladium or nickel powders (10 % on Vulcan XC-72 carbon) were used on a carbon cloth surface the power density of the

–2 NaBH4/O2 system was 33 mW cm lower than that obtained using Pt–Ni/C under similar conditions, which reached a value of 237 mW cm–2. A higher power density, up to 665 mW cm–2 (at 60 oC and with 1 mg Pt-Ni cm–2 loading and a Pd/C cathode

-2 1 mg cm ) was achieved in a NaBH4/H2O2 system using 10 wt. % NaBH4 (2.64 mol dm–3) in 20 wt. % (5 mol dm–3) NaOH aqueous solution and the oxidant concentration

–3 –3 was optimized as 2.0 mol dm H2O2 and 1.5 mol dm H2SO4 [64]. Among all the nickel based anode materials tested the best performance, obtaining highest power density, was that with electrodeposited Pd/C as cathode and Pt–Ni/C composite anode.

It should be taken into account that the cathodic compartment used H2O2 instead of O2,

25 Chapter 2. Literature Review and the use of that oxidant leads to higher current densities and power densities; but this will be analysed in the following section.

Zinc

Zn could be an appropriate anode catalyst for the DBFC since, in theory, it is a relatively poor electrocatalyst for hydrogen adsorption and reduction, it is a low cost material and it is suitable for energy storage [65]. Santos et al. [65] performed experiments such as cyclic voltammetry, chronoamperometry, and chronopotentiometry using a Zn disk of 0.5 cm2 surface area, a Pt mesh as a counter electrode and a solution

–3 –3 containing 1 mol dm NaBH4 in 4 mol dm NaOH as the anolyte. A volumetric mix

–3 ratio of 1:2:5 H2O2 in 1 mol dm HCl was used as the catholyte. A Nafion 117 membrane was used to separate the anode and cathode compartments. An electrode open circuit potential of –1.57 V vs. SCE was obtained, which, according to the

Pourbaix diagram [66], could correspond to the reaction:

(2.17)

The potential value was more negative than that obtained with other metals and offers the possibility of achieving a higher cell voltage. Cyclic voltammetry showed a peak corresponding to borohydride ion oxidation at –1.16 V vs. SCE, reaching a current density of 200 mA cm–2. At more positive potential values, a layer of Zn oxide is formed which covers the electrode surface and prevents the borohydride from being oxidised. During the reverse potential scan, the reduction of this layer occurs and the anode surface is reactivated allowing borohydride and/or its intermediates products to be oxidised. It was found that four electrons were transferred during the oxidation of

26 Chapter 2. Literature Review borohydride ions at this electrode. However, the hydrogen evolution from the borohydride hydrolysis at the Zn surface only takes place at very negative potentials (<

–1.9 V vs. SCE). A reaction mechanism proposed by Santos and Sequeira [65] consists of reaction (2.17) followed by:

(2.18)

– The BH4 /H2O2 cell voltage obtained by Santos et al. [65] was 2.14 V, which is lower

– than the theoretical (≈ 3 V) for a BH4 /H2O2 system. A stability test showed that the cell was able to operate for short periods of time (no more than 6 h), depending on the ohmic losses within the cell, which increase linearly as the current rises. Below cell voltages of 1.4 V, the discharge current dropped dramatically. The cell discharge curves showed a power density of 528 mW cm–2 for 0.8 s, a specific capacity of 1577 A h kg–1 and an energy density as high as 2799 W h kg–1. Polarization data showed anode limitations for a short time operation, caused by ohmic losses, particularly at high cell currents [65].

Palladium

Palladium is catalytically active towards both electrochemical oxidation of borohydride and its hydrolysis, yielding large anodic currents and high charge efficiencies at relatively low borohydride concentrations. The oxidation on palladium/carbon electrodes involves six electrons with 75 % efficiency as shown in the following reaction [31, 63]:

(2.19)

27 Chapter 2. Literature Review

However, this number of electrons can be reduced to 4 when a higher concentration of

– -3 BH4 (> 1 mol dm ) is used [31]. Celik et al. [67] carried out experiments in a direct

- BH4 /O2 (humidified air) flow cell using a MEA with a Pd/C anode and a Pt/C cathode divided by a Nafion membrane. A maximum power density of 27.6 mW cm–2 at a cell voltage of 0.85 V was obtained with a single cell of 25 cm2 active electrode area

o –3 –3 operating at 60 C (333 K), with 1 mol dm (3.7 wt. %) NaBH4 in 5 mol dm (20 wt.

%) NaOH. The highest power was obtained, unsurprisingly, when the highest anode catalyst loading was used, increasing by 34 % when the loading increased from 0.3 mg cm–2 to 1.08 mg cm–2 Pd/C. The current density can be considerably increased by changing the oxidant to H2O2.

Miley et al. [18] used a MEA arrangement consisting of Pd/C anode, Au/C cathode and a Nafion 112 cationic exchange membrane to construct a 500 W cell–stack of 15

2 NaBH4/H2O2 fuel cells with an active area of 144 cm per cell, resulting in a power density per cell of 231 mW cm–2. This value is comparably smaller than that obtained by Gu et al. [36], 680 mW cm–2, at 60 oC (333 K) using a Pd/C electrodeposited anode

-2 electrode and a Au/C (0.5 mg cm ) cathode electrode in a NaBH4/H2O2 fuel cell fed

–3 with a solution containing 10 wt. % (2.64 mol dm ) NaBH4, 5 wt. % NaOH and 5 wt.

% NH3OH in the anodic compartment and a solution containing 10 wt. % H2O2 in 5 wt.

% phosphoric acid. It is difficult to make a comparison between the results for the last two authors, due to the lack of information given by Miley et al. about the concentration of fuel and oxidant used. What it can be concluded is that Pd/C seems to have higher catalytic activity showing higher current density than any of the previously mentioned catalyst materials.

28 Chapter 2. Literature Review

Yi et al. [68] confirm that hollow nanospheres Pd/C exhibit higher electrocatalytic activity than carbon supported Pd solid nanoparticles for the borohydride oxidation, and can give a power density of 48.4 mW cm−2 at 54.8 mA cm−2 compared to 36 mW cm-2 at 45 mA cm-2 obtained using the Pd solid nanoparticles.

Some of the catalysts previously analysed, such as Au or Ag, showed an improved performance by alloying them with other metals, compared with the performance of the pure catalyst and the catalyst cost could be considerably reduced together with the amount of noble metal present in the alloy. That also happened in the case of Pd. Atwan et al. [69] investigated the borohydride oxidation on supported Pd and Pd-alloy nano- catalysts such as Pd-Ir, Pd-Ni, Pd-Au and Pd-Au (loading: 5 mg cm-2). All the catalysts were tested in a 5 cm2 single cell with a Pt cathode (loading: 4 mg cm-2). The anolyte (2

-3 -3 3 -1 mol dm NaBH4 in 2 mol dm NaOH) was fed at 50 cm min and the catholyte (O2)

200 cm3 min-1 at 2.7 atm. Among the catalysts tested, Pd-Ir was the most active; giving a current density of 50 mA cm-2 at a cell voltage of 0.5 V and at 298 K, compared to around 35 mA cm-2 obtained using pure Pd at the same conditions and same potential.

Therefore, by alloying Pd with Ir, the current density of the DBFC could be increased by 30 % compared to pure Pd. Considering that the Pd-Ir alloy showed the better

- performance as an anode catalyst for the BH4 /O2 fuel cell, and the high results obtained

- by Gu et al. [36], it would be interesting to perform a BH4 /H2O2 fuel cell using a Pd-Ir anode catalyst.

Iridium

Very few publications have been reported about the use of pure iridium as anode material for the borohydride oxidation. Kiran et al. [70] obtained a power density of 140

29 Chapter 2. Literature Review mW cm-2 using Ir/C as anode material for borohydride oxidation at 165 mA cm-2 at 80 oC (353 K). A MEA containing Ir/C (0.5 mg/cm2), Pt/C (2 mg/cm2) and a Nafion 117 membrane to separate anode and cathode. The concentrations of the anolyte and

-3 -3 -3 catholyte were 2.64 mol dm in 2.5 mol dm NaOH and 2.2 mol dm H2O2 in 1.5 mol

-3 dm H2SO4, respectively. By alloying the Ir catalyst with Rh, the peak power density of the cell increased to 270 mW/cm2 at a load current density of 290 mA/cm2.

Osmium and copper

Atwan et al. [42] reported cyclic voltammetry experiments on finely-divided Os particles supported on Vulcan carbon powder and Os-alloys (OS-Sn, Os-Mo and Os-V) for the direct oxidation of borohydride ions. A three-electrode cell with a graphite rod

–3 counter electrode was used. Using a solution containing 0.03 mol dm of NaBH4 in 2 mol dm–3 NaOH the oxidation potentials were between 0.1 V and 0.3 V vs. Ag/AgCl,

–2 KClstd with a peak current density of 40 mA cm for Os and values of around 40 mA cm–2, 20 mA cm–2 and 10 mA cm–2 for Os-Mo, Os-Sn and Os-V, respectively. It was also argued that the operating potentials of colloidal Os-alloys are more positive than those of Os colloids; therefore, alloying Os with Mo, Sn or V does not have any benefit.

In conclusion, borohydride ion oxidation was observed at colloidal Os and Os-alloys but low current densities were obtained.

Using a DBFC with a Cu anode, a Pt cathode and NRE212 membrane at room

–3 –3 temperature containing an anolyte of 2 mol dm NaBH4 in 2 mol dm NaOH, Zhi-fang et al. [71] found that the maximum current density and maximum power density were

235 mA cm–2 and 46.14 mW cm–2, respectively. A stable cell voltage of around 0.6 V was obtained for 50 h at a current density of 20 mA cm–2. Copper is active against the

30 Chapter 2. Literature Review borohydride oxidation, but copper oxide is formed at negative potentials, very close to that of borohydride oxidation. Thus copper is not as good as Pt, Pd or Au for borohydride oxidation. However, as it has been previously mentioned in this section,

Au-Cu alloys have shown better performance for the DBFC than the pure Au and obviously than pure Cu.

Other catalyst materials

AB5 and AB2-type hydrogen storage alloys have also been identified as anode catalysts for DBFCs [72, 73]. These metallic materials have the ability to absorb and release significant amounts of hydrogen gas. In the AB5 alloys, A is an hydride forming metal, usually a rare earth metal (e.g., La, Ce, Nd, Pr, Y or a mixture (mischmetal)), and B is a non-hydride forming element, such as Ni, which can be doped with other metals (e.g.,

Co, Sn or Al) to improve the stability or to adjust the equilibrium hydrogen pressure and temperature required to charge/discharge hydrogen [73]. In the AB2 alloys, A represents a large group of alloys containing Ti, Zr or Hf, and B represents a transition metal (e.g.,

Mn, Ni, Cr or V). Higher power densities were achieved with the AB5-type alloys such

-2 -2 as MmNi3.55Al0.3Mn0.4Co0.75 (5mg cm ), with which 150 mW cm was obtained,

-2 compared with 70 mW cm , obtained when Zr0.9Ti0.1V0.2Mn0.6Cr0.05Co0.05Ni1.2 (AB2- type) (5mg cm-2) was used as anode material. The same cathode material, 60 wt. % Pt/C

(1mg cm-2 of Pt), the same operation conditions and solutions were used in both cases

[47]. The higher power density obtained using AB5-type alloys compared to AB2-type alloys is probably due to their lower catalytic activity towards the borohydride hydrolysis and high capability to oxidise the hydrogen generated [74].

31 Chapter 2. Literature Review

Hydrogen storage alloys are anticipated to slow hydrogen evolution in DBFCs since they can absorb large quantities of hydrogen gas. Wang et al. [74] reported that the

LmNi4.78Mn0.22 (AB5-type hydrogen storage alloy, where Lm is a La-rich mischmetal) hydrogen storage alloy exhibits both high electrochemical catalytic activity for hydrogen generation and the oxidation of borohydride. Although this appears to lead to an inefficient anode material the authors suggested that an electrochemical surface treatment of the alloy can help to decrease the hydrogen evolution. The alloy was modified by the addition of Si and by heat treatment and used as an anode catalyst in a

DBFC. The rate of hydrogen generation was decreased by modifying the alloy resulting in and enhanced fuel utilization efficiency of 4.5 times from 21.4 % to 95.3 %. however, the electrochemical catalytic activity of the anode was appreciably lower [75].

Various AB5- and AB2-type hydrogen storage alloys have been employed as anode catalysts in DBFCs [76-81]. Lee et al. [34] reported that the use of ZrCr0.8Ni1.2 (AB2- type alloy) led to hydrogen generation from borohydride through a stepwise mechanism followed by the electrochemical oxidation of hydrogen to generate electricity.

Choudhury et al. [47] reported a power density of 150 mW cm–2 at a cell voltage of 0.54

o –2 V while operating at 70 C (343 K) using MmNi3.55Al0.3Mn0.4Co0.75 (5 mg cm ) as the anode catalyst and 60 wt. % Pt on carbon as the cathode catalyst, with a platinum

−2 –3 loading of 1 mg cm . A solution containing 10 wt. % (2.64 mol dm ) NaBH4 in 20 wt.

% (5 mol dm–3) NaOH was used. Li et al. [82] obtained a higher power density, 290 mW/cm2, with a 5 cell stack of 67 cm2 active area. A power of 110 W was achieved starting at room temperature and reaching 60 oC (333 K) during operation. The cell consisted of a Pt/C air cathode, a Nafion membrane (NRE-211) and a mixture of surface treated Zr–Ni laves phase alloy AB2 (Zr0.9Ti0.1Mn0.6V0.2Co0.1Ni1.1) and Pd/C as the

32 Chapter 2. Literature Review anode catalyst. Other materials such as Ru or Co have been reported as good catalysts for the borohydride hydrolysis and those would be appropriate for an IBFC rather than a

DBFC [83, 84].

2.1.2. Cathode

Two oxidizers have been used extensively for the cathodic reaction in DBFCs: O2 and

H2O2. Hydrogen peroxide is a good alternative for anaerobic applications, namely, underwater vehicles and space [4, 18]. The reduction of H2O2 is a faster process than that of oxygen, which enhances the power density and the cell efficiency [36]. Coupling these oxidants with the borohydride reaction provides different theoretical cell potentials depending on the electrolyte medium as shown below:

1- Oxidant: O2 Alkaline electrolyte

0 E = 0.4 V vs. SHE (2.20)

2- Oxidant: H2O2 Alkaline electrolyte

0 E = 0.87 V vs. SHE (2.21)

3- Oxidant: H2O2 Acid electrolyte

0 E = 1.77 V vs. SHE (2.22)

33 Chapter 2. Literature Review

Comparing the cell voltages using different oxidants and electrolyte mediums,

NaBH4/H2O2 fuel cells exhibit higher cell voltages in acidic electrolyte, Ecell  3 V, which represents a predicted energy density of 17,060 W h kg−1 as is shown in Table

1.1 [4]. From the operational point of view, the fact that all the components are liquid presents advantages in terms of storage, pumping requirement and heat removal.

The preceding section highlights the vast effort directed at finding suitable anode catalysts for DBFCs. In this section the cathode catalyst and its influence on fuel cell performance will be analysed [36, 41, 43-45]. The choice of cathode catalyst for O2 reduction has been rather limited; most investigators have used Pt. Alternatives, such as

Pt/C, Au/C, Ag/C, MnOx/C and MnOx-Mg/C, were investigated by Chatenet et al. [41].

The authors suggested that Pt/C, Ag/C and Au/C should not be used as DBFC cathode

− catalysts, since they exhibit electrocatalytic activity towards BH4 oxidation. This reaction competes for the active sites on the catalyst surface, leading to deterioration in the cell performance. Chatenet et al. maintain, however, that MnOx/C and MnOx-Mg/C are unaffected by the presence of borohydride and are, therefore, appropriate as cathode

– catalysts in BH4 /O2 fuel cells [17].

– In BH4 /H2O2 fuel cells, Pt, Au, Pd, Ag, Raney Ag, Pd/Ir, Pd-Ru, and Pd-Ag have been studied as cathode catalysts for H2O2 reduction [41, 43, 44, 85, 86]. A good cathode catalyst will reduce the hydrogen peroxide to produce water as shown in reactions

(2.21) or (2.22) and will not decompose it to produce oxygen according to the following reaction [36]:

(2.23)

34 Chapter 2. Literature Review

Pt in the cathode typically yields excellent performance in terms of power density; however, it tends to decompose peroxide very rapidly [18]. Both Au and Pd based cathode catalysts have exhibited good performance [36, 45], minimizing oxygen evolution from the hydrogen peroxide decomposition and providing high power densities. A peak power density of 680 mW cm–2 at 60 oC (333 K) was achieved in a

– BH4 /H2O2 fuel cell with a Pd anode catalyst, prepared by applying a 1:1:20 mixture of

Nafion-Pd-methanol on a carbon cloth, which acts as a diffusion layer, and an Au cathode catalyst (0.5 mg cm-2) in an anolyte solution containing 4.75 mol dm-3 (18

-3 wt. %) NaBH4 in 4.25 mol dm (17 wt. %) NaOH [36].

-2 The performance of a BH4/O2 fuel cell with an Au/C (2 mg cm ) anode in combination with different cathode materials is shown in Figure 2.4 together with the comparison of the same cathodes but in a BH4/H2O2 fuel cell with a Pd/C as anode. The curves in the figure show that the peak power density varied from 33 to 35 mW cm-2, when Ag/C and

Ni/C were used, respectively, and 65 to 72 mW cm-2 when iron tetramethoxyphenyl prophyrin (FeTMPP) and Pt/C, respectively were used [39].

Other cathode materials based on cobalt phthalocyanine have shown considerable activity for oxygen reduction without been influenced by the concentration of sodium borohydride giving up to 90 mW cm-2 at a discharge current density of 175 mA cm-2 in a simple DBFC [77]. The figure also shows that higher current densities can be obtained

-2 by using H2O2 as the catholyte, varying the power density from 90 to 171 mW cm for

Ag and Pd, respectively at a discharge current density of 180 mA cm-2 [36].

35 Chapter 2. Literature Review

200

180 a) 160

140

2 -

120

b) / mW cm mW /

100 P 80 c) 60

40 d) Power density, density, Power e) 20 f) g) 0 0 50 100 150 200 250 300 350

Current density, j / mA cm-2

Figure 2. 4 Power density (P) vs. current density curves measured in a BH4/H2O2 fuel cell at 25 oC using a Pd/C anode of 25 cm2 active area and different cathode materials [43]: a) Pd, b) Pt, c) Ag; and in a BH4/O2 fuel cell at 85 oC, using an Au/C anode of 4 cm2 active area (2 mg Au cm2) and different cathode materials (2 mg cm2): d) Pt/C, e) FeTMPP, f) Ni/C and g) Ag/C. Membrane: Nafion 117. Fuel: 5 wt. % (1.32 mol 3 -3 dm ) NaBH4 in 10 wt. % (2.5 mol dm ) NaOH [2, 77, 87].

2.1.3. Membranes

The ion-exchange membrane prevents borohydride ions from being in contact with the cathode, which can be active for borohydride decomposition. The presence of borohydride ions can result in the deactivation of the cathode catalyst, probably due to the formation of a borate layer on the surface. Deactivation can be minimised by using a highly selective membrane combined with the use of a cathode catalyst that is inactive towards borohydride decomposition (e.g. MnO2), as well as an optimized concentration of fuel [21]. High concentrations of borohydride can lead to a high degree of crossover

[88] through the membrane towards the catholyte side due to concentration gradient.

36 Chapter 2. Literature Review

Two types of membrane can be used in fuel cells: the anion-exchange and the cation- exchange membranes. In the DBFC, an anion-exchange membrane will transport hydroxyl ions from the catholyte to the anolyte compartment maintaining the alkaline pH of the anolyte at sufficiently high levels to ensure the stability of the borohydride ions, as can be seen in Figures 2.5(a) and 2.5(c). If a cation-exchange membrane is used, sodium ions will migrate from the anode to the cathode compartment during operation to maintain charge balance, as shown in Figures 2.5(b) and 2.5(d). This is the drawback of using a cation membrane, since the Na+ exchange will cause the alkalinity of the anolyte to decrease during operation; thus, the borohydride will become unstable.

The majority of DBFCs use a cation-exchange membrane because they severely hinder

NaBH4 crossover, in particular the Nafion membrane. This membrane is also selected by virtue of its mechanical and chemical stability in a strongly alkaline environment, although its long term stability is still questionable, particularly at high temperatures

[89]. Improved performance is obtained with thin membranes, which is due to the correspondingly lower ohmic resistance. Thin membranes, however, exhibit higher rates of borohydride ion crossover, leading to larger deviations in the OCV from the theoretical value [17, 90].

According to Raman and co-workers [80] the use of a Nafion (R)-961 membrane helps

– to prevent the crossover of BH4 [90]. Initial results indicated that the fuel cell exhibited a maximum power density of 10 mW cm–2 at an operating voltage of 0.77 V with an oxidant utilization of about 80% at 25 °C (298 K). Higher power densities were obtained using a non-commercial membrane made from polyethylenetetrafluoroethylene (ETFE-g-PSSA); the power density 103 mW cm-2 and

-2 -3 -3 the peak current 475 mA cm of a DBFC (1.32 mol dm NaBH4 in 2.5 mol dm

37 Chapter 2. Literature Review

NaOH, 0 bar oxygen and 85 oC (258 K)) was also higher than that using a Nafion 117

membrane, 72 mW cm-2 and 350 mA cm-2.

Anode 8e- Anode 8e-

Cathode Cathode

8OH-

O2 O2 Fuel Fuel flow flow flow flow Na+

Anionic exchange membrane Cationic exchange membrane

a) b) Cathode Anode Anode Cathode 8e- 8e-

u 8OH- e Fuel Fuel l flow flow O 2 O2 f flo flow Na+ l w

o Anionic Cationic w exchange exchange membrane c) membraned)

Figure 2. 5 Different configurations of the direct borohydride fuel cell (DBFC): MEA configuration where the electrodes are separated by: a) an anionic exchange membrane and b) a cationic exchange membrane. Oxygen reduction in an alkaline media in a flow system where the electrodes are separated by the electrolytes and by c) an anionic exchange membrane and d) a cationic exchange membrane.

38 Chapter 2. Literature Review

The open circuit voltage, nevertheless, was found to be higher when a Nafion membrane was used [91]. Ma et al. [92] recently reported the use of a 200 μm polyvinyl alcohol hydrogel membrane (PHME) in a DBFC with a Ni-Pt/C anode and a gold sputtered cathode using oxygen, humid air and hydrogen peroxide as oxidants. They compared their results at the same experimental conditions using a Nafion 212 membrane, and obtained lower power densities with the PHME membranes: 218 mW cm–2 and 176 mW cm–2, respectively at 60 oC (333 K), probably due to a higher crossover rate. The results, however, were comparable and even slightly better than with Nafion membrane when using oxygen: 242 mW cm–2 at 60 oC (333 K). The

– stability test showed that the BH4 /H2O2 fuel cell discharged a current density of 50 mA cm–2 for 100 hours at ambient temperature.

Membranes tend to be expensive and they complicate the cell design, and there is always a slow rate of undesirable crossover, which requires frequent refurbishment of the solutions. Experiments with an undivided DBFC have been reported by Feng et al.

[17] who demonstrated that MnO2 could be used as a cathode material in a DBFC without a membrane or using a conventional alkaline membrane instead of an expensive electrolyte membrane (i.e. Nafion). This is due to the absence of chemical reactions or

– crystalline transformation at the MnO2 surface in contact with BH4 . A cell voltage of

0.6 V with a current density between 1 and 5 mA cm–2 was obtained by using a a membrane-less system with a dispersed gold catalyzed anode (7.4 – 8 electrons

–3 –3 interchanged) and a solution containing 1 mol dm KBH4 in 6 mol dm KOH and a

MnO2 catalyzed air cathode. Better performance was achieved by Verma et al. [90] who, in contrast with Feng et al. [17], used a Pt/Ni anode in a flowing alkaline

39 Chapter 2. Literature Review electrolyte and obtained 19 mW cm–2 with a current density of 39 mA cm–2 using 1 mol

–3 –3 dm NaBH4 in 3 mol dm KOH and no membrane.

2.1.4. Cell performance

The cell voltage of a fuel cell or any electrochemical cell deviates from the theoretical as a consequence of the following parameters represented in Figure 2.6:

1) Activation polarization at the electrodes, representing the energy barriers to

charge transfer reactions away from (dynamic) equilibrium at the OCV.

2) (Ohmic) resistances to charge transport in the electrolyte, membrane/separator,

electrodes and current-collecting plates, together with resistances due to

imperfect contacts between layers of active materials electrode support and

current collectors.

3) Concentration polarisation arising from limitations in the mass transport of

reactants to the reaction sites (typically catalyst surfaces).

4) Reactant crossover between electrode compartments.

Depending on the cell materials, operating conditions and cell chemistry, different overpotential losses will dominate. Generally, activation losses dominate at low cell current densities, ohmic losses dominate at intermediate current densities and concentration polarization losses dominate at high current densities. Internal currents and fuel crossover losses are apparent from the OCV, which is typically lower than the theoretical standard value, often appreciably. Considering all the parameters mentioned above, the cell voltage can be expressed according to the following equation [93]:

40 Chapter 2. Literature Review

| | | | ∑ ( ) | | | | (2.24)

where E is the cell voltage, Eo is the predicted open-circuit cell voltage, a ( c ) cell cell hact hact is the activation polarisation at the anode (cathode), (IRe)k is the ohmic drop (electronic and ionic) across component k, and a ( c ) is the concentration polarisation at the hconc hconc anode (cathode).

Ecell

0 Ecell ,theoretical

Internal currents/fuel crossover / Development of rest potential 0 Ecell a c Activation losses: hact + hact

Ohmic losses: ∑ å(IR)k k

a c Concentration losses: hconc + hconc

Cell current density jcell

Figure 2. 6 A parallel-plate electrode unit cell showing the location of elements that contributes to the overall cell voltage drop [2].

The use of the term ‘activation polarisation’ rather than ‘activation overpotential’ is quite deliberate since the anode of a DBFC operates at a mixed potential, while a mixed potential can also be established at the cathode when H2O2 is used as the oxidant, due to the occurrence of competing anodic and cathodic reactions at both electrodes. The influence of each of these terms on the cell voltage can be seen in Figure 2.6, where the cell voltage has been plotted versus the current density. The three areas mentioned above: activation polarizations, ohmic drop and mass transport limitations can be

41 Chapter 2. Literature Review clearly seen. Figure 2.7 shows the influence of these different components of the cell on the cell voltage in a schematic diagram of an electrochemical cell:

E Ion exchange membrane IR 0 circuit E anode 0 Eanode = E anode +| |

IRanolyte a A Ecell t n IRmembrane h o o d IRcatholyte d e 0 Ecathode = E cathode + | | e 0 E cathode IRcircuit

0 Distance, x

Figure 2. 7 Schematic diagram of a parallel-plate electrode cell showing the elements that contributes to the overall cell voltage drop. The voltage across the electrodes, electrolyte and membrane is plotted versus the distance between the two parallel electrodes [94].

2.1.5. Hydrogen evolution and the use of inhibitors

Hydrogen evolution is a major obstacle to commercial DBFC development. The hydrogen generation during the fuel cell operation not only reduces the amount of fuel available to be used by the fuel cell but also can generate safety problems. As the hydrolysis reaction rate depends on the anode catalysts material and the borohydride concentration, investigations have been carried out to measure the hydrogen generation rate on several anode materials.

Liu et al. [31] measured the volumetric hydrogen generated at various anodic currents and sodium borohydride concentrations using Ni, Pd/C and Pt/C as anode catalysts.

42 Chapter 2. Literature Review

Wang et al. [95] measured the hydrogen generation rate and the anodic current at different electrode potentials and sodium borohydride concentrations using Ni, Pt/C,

–3 Au/C and Cu. A comparison of their results obtained for 0.5 mol dm NaBH4 in 6 mol dm–3 NaOH and in 2 mol dm–3 NaOH, respectively is shown in Figure 2.8.

20 a)

18 -1

min 3 14

6 b) 4 d) c) Hydrogen evolution ratecm / evolution Hydrogen 2 e) g) 0 f) 0 100 200 300 400 500 600 Current density, j / mA cm-2

Figure 2. 8 Hydrogen evolution rate vs. anodic current. a) 1 g Ni (Inco type 210, particle size 0.5–1.0 μm ), d) 10 wt. % Pd/C, g) 5 wt. % Pt/C: 0.5 mol -3 -3 dm NaBH4 in 6 mol dm NaOH at room temperature [31]; b) Ni (Inco type 255 particle size 2.2–2.8 μm), c) 20 wt. % Pt/C e) 20 wt. % -3 -3 Au/C, f) Cu: 0.5 mol dm NaBH4 in 2 mol dm NaOH [95].

In all cases, the rate of hydrogen generation initially decreased as the current was increased but then increased dramatically at high currents, with the most dramatic increase observed when a Ni electrode was used. The results obtained by the different authors using Ni electrodes were different; however, the particle sizes variation between the cases should be considered. Faster rates of hydrogen generation were observed with

43 Chapter 2. Literature Review small particle sizes, due to the larger active area surface. The lowest rates of hydrogen generation at high current densities were observed on Pd and Pt electrodes, curves d) and g) respectively, however, at the open circuit potentials the hydrogen generation on these metals is also significant. When the Pt/C electrode was used, Wang et al. [95]

3 –1 obtained a rate that was almost 2 cm H2 min faster than that observed by Liu et al.

[31] for the same borohydride concentration and temperature conditions. This could have been due to the use of a higher catalyst loading and lower concentration of sodium hydroxide.

It was also demonstrated in both studies that higher initial concentrations of borohydride lead to faster hydrogen generation rates. Liu et al. [31] reported that at a borohydride concentration below 1 mol dm-3 in alkaline solutions it was possible to obtain a quasi-8e- reaction on Pt anode. It was suggested that a maximum of 3 hydrogen

- - atoms from the BH4 ions can be simultaneously adsorbed, and that the additional 2 e in

- the quasi-8e reaction might be result of H2 electrooxidation. In a subsequent study, Li et al. [96] suggested that coating a thin Nafion film on the catalyst surface can decrease the hydrogen generation rate. If the Nafion loading was too high, however, the access of the fuel to the active sites was limited. The optimum Nafion content was found to be less than 25 wt. %. The hydrogen generation rate was also reduced by decreasing the temperature, which carried a penalty in terms of the cell performance.

Thiourea (TU) and tetraethyl ammonium hydroxide (TEAH) have been proposed as inhibitors for the borohydride hydrolysis reaction [19]. There is a diversity of opinions about the use of TU as an inhibitor for the hydrolysis of borohydride in a DBFC.

Demirci [97] suggested that the adsorbed molecules of TU can block the active sites on

44 Chapter 2. Literature Review the anode catalyst. This limits the formation of M-H bonds and, therefore, the hydrogen generation rates, but it also limits the catalyst active sites available for the borohydride oxidation. Jamard et al. [98] concurs with Demirci, arguing that although TU reduces the hydrogen generation rate, it also decreases the cell power density. Cyclic voltammogram studies, conducted using a three-electrode cell with a Pt disk electrode

–3 –3 and a solution containing 0.1 mol dm NaBH4 in 1 mol dm NaOH in the absence and in the presence of 10–3 mol dm–3 and 10–4 mol dm–3 TU, demonstrated that the current density peak tends to decrease with an increasing concentration of TU. Polarization curves showed that when TU was added the OCV was lower than the theoretical value, and that the power density peak was reduced by 50 % (90 mW cm–2 in the presence of

10–4 mol dm–3 TU) [98].

Although all authors agree that TU inhibits the ionization and liberation of hydrogen,

– and does not affect the BH4 electrooxidation; some authors assert that the performance of the fuel cell can be improved when adding TU [99]. Figure 2.9 shows cyclic voltammograms from Martins et al. [19] in the presence and in the absence of TU using a three-electrode cell consisting of Pt on a Ni mesh as the working electrode, and

Ag/AgCl and a Pt mesh as the reference and counter electrodes, respectively. A solution

–3 –3 of 0.03 mol dm NaBH4 in 2 mol dm NaOH in the absence and in the presence of different concentrations of TU was used. In the absence of TU, three oxidation peaks were observed: the peak a1 is due to oxidation of the hydrogen generated through the hydrolysis of borohydride (reactions (2.8) and (2.11)); the peak a2 is due to the direct

– oxidation of BH4 (reaction (1.5)); and the third peak in the reverse cycle, c1 (not shown

– for clarity), is due to the intermediate product (BH3OH ) oxidation (reaction (2.11)). In the presence of TU it was observed that the peaks a1 and c1 disappear, which means that

45 Chapter 2. Literature Review

– TU inhibits the oxidation of H2 and possibly the catalytic hydrolysis of BH4 in conjunction with the oxidation of the intermediate species (reactions (2.9) and (2.10) do not take place in this case). Only an oxidation peak at the same potential as the

– oxidation of BH4 on Pt (–0.2 V vs. Ag/AgCl) appears.

The effects of different concentrations of TU on the oxidation of borohydride were studied by Martins et al. [19] by adding TU at 5 × 10-3 mol dm-3, 1.5 × 10-2 mol dm-3, 2

× 10-2 mol dm-3 and 3 × 10-2 mol dm-3 to a solution containing 2.6 × 10-2 mol dm-3

-3 NaBH4 in 3 mol dm NaOH using Pt rod as a working electrode. Cyclic voltammograms showed that when TU was in the solution at a concentration of 3 × 10-2 mol dm-3 and without sodium borohydride, the oxidation peak was higher than that

- obtained in the presence of BH4 ; which means that the TU could be oxidised at the same potential as the borohydride ions, although further investigation needs to be carried out to confirm it. Higher current was obtained when the concentration of TU was higher (2 × 10-2 mol dm-3 compared to 1.5 × 10-2 mol dm-3). The authors concluded that the addition of TU can improve the performance of the cell, decreasing the hydrogen generation and increasing the charge efficiency; but there is an optimum concentration above which the TU impacts negatively on the cell performance, decreasing the current density associated with the borohydride oxidation.

–3 –3 –3 Celik et al. [99] showed that adding 1.6 × 10 mol dm TU to 1 mol dm NaBH4 in 20 wt. % (5 mol dm–3) NaOH solution, the power density increased from 14.4 mW cm–2 to

15.1 mW cm–2, using a Pd/C electrode, due to the decreased rate of hydrogen evolution.

Atwan et al. [42] carried out cyclic voltammograms on supported colloidal Os in a three-electrode cell with a graphite rod and Ag/AgCl, KClstd as the counter and

46 Chapter 2. Literature Review reference electrodes, respectively. Figure 2.9 also shows cyclic voltammograms from electrolytes containing 2 mol dm–3 NaOH, 2 mol dm–3 NaOH with 1.5 × 10–3 mol dm–3

–3 –3 –3 –3 TU, and 0.03 mol dm NaBH4 in 2 mol dm NaOH with 1.5 × 10 mol dm TU on

– 10 wt. % Os. The oxidation potential of TU and BH4 appear to be between 0.1 V and

0.3 V vs. Ag/AgCl, KClstd, meaning that they both oxidise at the same potential and the

– oxidation wave is a mixture of both the TU and BH4 oxidation currents.

40 a 1 a 400 1

30 -2

-2 1)

300 cm





/ mA / j

/ 20 j 200 aa2 2 2)

10 100 3)

4) density, Current Current density, density, Current

0 5) 0

-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 Electrode potential, E vs. Hg/HgO / V

Figure 2. 9 Cyclic voltammogram on colloidal 10 wt. % Os supported on carbon support (Vulcan XC-72) at 100 mV s-1 and 298 K [42]: 1) 0.03 mol -3 - -3 -3 -3 dm BH4 + 2 mol dm NaOH + 1.5 × 10 TU, 2) 2 mol dm NaOH + 1.5 × 10-3 TU and 3) 2 mol dm-3 NaOH; together with CV of 0.03 -3 -3 mol dm NaBH4 + 2 mol dm NaOH on Pt disc electrode (1 mm diameter): 4) in the absence of TU and 5) in the presence of 1.5 × 10-3 mol dm-3 TU [32].

47 Chapter 2. Literature Review

TEAH, in contrast to TU, does not eliminate the voltammetric responses due to hydrogen evolution. Instead, it shifts the oxidation peak of the borohydride ions by +0.3

V, which could be due to adsorption of the TEAH ion on the Pt electrode [32]. The fact that the presence of TEAH shifts the potential to more positive values implies an increase in the overpotential required for the borohydride oxidation, and thus the use of

TEAH is not beneficial for the borohydride oxidation or to inhibit its hydrolysis.

Further investigations are necessary to find a more effective additive to inhibit the hydrolysis of borohydride and improve the performance of the cell and its coulombic efficiency. Surfactants have been shown to influence other electrode reactions. For example, Dang et al. [100] demonstrated that the presence of the cationic surfactant cetyltrimethylammonium bromide (CTAB), on the surface of an acetylene black electrode, significantly decreased the overpotential of oxygen reduction and increased the peak reduction current. There are very few publications reporting the role of surfactants in the modification of electrode surfaces, making their exact role on electrode surfaces difficult to understand [101]. It is clear, however, that the presence of particular surfactants in solution can significantly alter the voltammetric response of an electro-oxidation reaction. For example, He et al. [102] increased the oxidation of estradiol on a nano-Al2O3 modified glassy carbon (GC) electrode by adding CTAB to the solution. The peak current showed a complex dependence on the concentration of surfactant, firstly increasing at higher surfactant levels, remaining constant when the surfactant concentration was further increased and finally decreasing due to the formation of micelles, which can block access to the electrode and lower the rate of electron transfer. Due to the complex interaction between the electrode and electrolyte

48 Chapter 2. Literature Review additives, the surfactant can increase or diminish hydrogen generation during borohydride oxidation. Again, this has been rarely studied.

Ferreira et al. [103] showed that the hydrogen generation and hydrogen storage capacities increased up to 6 wt. % on the addition of 0.25 wt. % of the organic polymer carboxymethyl cellulose (CMetC). The authors reported using a solution containing 10 wt. % NaBH4, 7 wt. % NaOH, 83 wt. % H2O and 0.25 wt. % CMetC, in a stainless steel batch reactor of 0.229 dm3, with a conical bottom shape. A Ni–Ru catalyst was used over 250 times at 45 oC. Therefore, it is believed that the presence of surfactants may alter the borohydride oxidation and its hydrolysis reaction. Therefore, the use of nonionic and anionic surfactants, such as FC4430, Triton X-100, Zonyl FSO, S-228M and sodium docdecyl sulphate (SDS) will be studied in this work. Appendix I summarizes the structure and properties of each surfactant used in the present study.

2.1.6. Influence of operational conditions in a DBFC a) Effects of temperature on DBFC performance

One advantage of borohydride fuel cells is their capability to provide energy at, or near, room temperature, which makes them suitable for portable applications. There is evidence, however, that performance parameters such as power density and current density improve at elevated temperatures [63, 104]. Raising the temperature presents some advantages, such as: high mass transport rate of the reactants, faster kinetics of borohydride electrooxidation and higher ionic conductivity of the electrolytes and within the membrane. But high temperatures also enhance the crossover rate and the

– hydrolysis of BH4 , which means a lower fuel utilization and cathode deterioration. The elevated temperatures may also cause dehydration of the membrane, increasing the membrane resistance and severely deteriorating the cell performance [64, 88]. Ma et al.

49 Chapter 2. Literature Review

[64] carried out experiments in a borohydride fuel cell comprising a Nafion® 212 membrane, a Ni + Pd/C composite anode, and 1 mg cm−2 Pt/C in the cathode. The fuel

–3 –3 was 5 wt. % (1.32 mol dm ) NaBH4 and 10 wt. % (2.5 mol dm ) NaOH aqueous solution, with a flow rate of 5 dm3 min−1. The oxidant was humidified oxygen or humidified air with a flow rate of 0.15 dm3 min−1. The power density increased from 77 mW cm−2 to 167 mW cm−2 only by increasing the temperature from 28 oC to 60 oC. The increase in temperature, however, also caused an increase in the hydrogen evolution rate, which was not quantified [105].

b) Effects of borohydride concentration on DBFC performance

According to the Nernst equation, the OCV would be expected to increase with an

– increased concentration of BH4 . Experiments have shown that an increase in the

NaBH4 concentration improves the performance of the anode but also increases the cathode polarization [48], probably as a result of increased borohydride crossover. This dependence of the power density with the NaBH4 concentration became smaller at high concentrations of borohydride ions [31]. Cheng et al. [88] performed experiments

–3 increasing the concentration of NaBH4 from 3 wt. % to 10 wt. % (0.79 mol dm and

–3 –3 2.64 mol dm ) in 10 wt. % (2.5 mol dm ) NaOH in a NaBH4/O2 fuel cell with a

Nafion® 117 membrane, 2 mg cm−2 Au/C anode, 2 mg cm−2 Pt/C cathode. The peak power density increased about 30 % when the concentration of borohydride increased from 3 wt. % to 5 wt. %, but did not increase appreciably with further increases in the concentration; an increase of 10 % was achieved by doubling the borohydride concentration to 10 wt. %. This is not surprising, considering that the Au loading was kept constant.

50 Chapter 2. Literature Review

In summary, by increasing the concentration of borohydride: the mass transport limitations alleviate, leading to higher power densities and higher current densities; but the borohydride crossover increases together with the hydrogen generation [88, 106]

[107], which deteriorates the cathode performance, decreases the fuel utilization and reduces the OCV [88]. Hence, the fuel concentration should be optimised, taking into account the cell performance and fuel cost. Likewise, the concentration of NaOH should be considered because it will influence the conductivity and alkalinity of the solution

(ohmic losses, particularly through the electrolytes and membrane, are high). By increasing the concentration of NaOH, the fuel crossover rate decreases as a result of increased numbers of migrating Na+/OH– ions blocking the passage of borohydride ions through the membrane and improving the performance of the cell [88]. However, excessive sodium hydroxide concentrations increase the viscosity of the solution and reduce the mobility of Na+ ions [6]; also hindering the movement of borohydride ions towards the reaction sites and partially inhibiting the borohydride oxidation [48]. In agreement with that, Cheng et al. [88] obtained 10 % higher power density and current density when increasing the NaOH concentration from 5 wt. % to 10 wt. % (1.25 mol dm–3 to 2.5 mol dm–3) but it decreased for a NaOH concentration of 20 wt. % (5 mol dm–3).

c) Effects of fuel flow on DBFC performance

The fuel/oxidant flow rate is another factor that affects the DBFC performance.

Experiments with flow rates of 10 cm3 min–1, 20 cm3 min–1, 90 cm3 min–1 and 150 cm3 min–1 were carried out by Celik et al. [67], who used a 25 cm2 cell containing an MEA composed by an Pd/C anode, a Pt/C cathode and a Nafion 117 membrane. Analysing the polarization curves it was concluded that enhancing the oxidant flow rate, a small

51 Chapter 2. Literature Review improvement in power density was obtained (8.5 mW cm–2 and 10.1 mW cm–2 at 10 cm3 min–1 and 150 cm3 min–1, respectively). Cheng and Scott [88] carried out experiments at flow rates from 10 cm3 min–1 to 200 cm3 min–1 using Au and Pt supported on carbon catalysts assembled in a 4 cm2 active area MEA. They suggested that a higher flow rate improves the performance of the cell by promoting the mass transport and reducing possible channel blocking and product accumulation [88].

Duteanu et al. [38] argued that the effect of the anolyte flow rate on the fuel cell performance is small and it is more economical to use low flow rates. Kim et al. [108] increased the power density of the cell by 21 % by increasing the air flow rate from 5 ×

103 to 104 cm3 min–1 and keeping a constant fuel flow rate of 108.4 cm3 min–1, indicating that the performance of the cell was also significantly affected by the air flow rate in the cathode and the overpotential of the cathode reaction.

Higher fuel rates will undoubtedly improve the mass transfer characteristics and lead to more uniform reactant distributions in the cell (important for avoiding dead zones of low reactant concentration and, ultimately, regions of high overpotential). A higher flow rate may also improve mass transport by removing the evolved hydrogen gas more effectively and replace them with fresh solution. There will be an optimal flow rate, however, above which improvements in performance are not compensated by the additional pumping power required. Higher flow rates also place additional stresses on the electrode materials, sealing materials and piping, increasing the risk of electrolyte leakage.

52 Chapter 2. Literature Review

2.1.7. Engineering aspects of direct borohydride fuel cells

Different flow cell designs have been investigated for the DBFC in order to improve its performance. Half-cell designs, such as the three electrode cell shown in Figure 2.10, are typically used to calculate the catalytic activity of the electrodes, i.e. kinetic parameters and constant rates together with its mechanical stability.

W RE CE E

NaBH4 in alkaline solution

Pt cathode counter Luggin capillary electrode

Alkaline solution

Working electrode Cation exchange membrane

Figure 2. 10 Typical three-electrode glass cell to carry out fundamental voltammetric studies of the borohydride oxidation reaction such as kinetic rate constants, exchange current densities and mass transport characteristics [109].

This electrochemical cell comprises working, counter and reference electrodes, with the counter electrode placed in a different compartment separated by a membrane [109].

53 Chapter 2. Literature Review

In the DBFC, the ion-exchange membrane can be sandwiched between the anode and cathode forming the membrane electrode assembly (MEA), as shown in Figures 2.5(a) and 2.5(b). Otherwise, the anode and cathode can be placed facing each other, as it is shown in Figures 2.5(c) and 2.5(d), leaving a space between electrode and membrane for the electrolyte to flow through. In both cases the fuel, sodium borohydride in alkaline solution, and the oxidant, either hydrogen peroxide solution, air, or pure oxygen, are pumped from separate tanks to the anode and cathode compartments, respectively. The fuel and oxidant pass through the flow field plates, normally made of graphite or stainless steel, towards the anode and cathode and react to exchange a maximum of 8 electrons [110]. The hydrogen gas, primarily from borohydride hydrolysis, will pass through the flow fields together with the borohydride solution, decreasing the ionic conductivity of the electrolyte and the effective diffusion of the reactants.

Serpentine and parallel flow channels are the most common flow field designs used in direct liquid fuel cells. The serpentine channel is typically used in the anode, since it facilitates mass transport of the fuel and increases the fuel velocity, generating even current and potential distribution; compared to that in the parallel flow channel at the same solution flow rate. Higher liquid velocity enhances the mass transfer of borohydride from the flow channel to the gas diffusion layer, thereby improving the cell performance [111]. A long narrow channel will increase the performance but also the pressure losses; at low flow rates, the gas bubbles might hinder the liquid flow, which will decrease the cell efficiency. A serpentine flow field can reduce the degree of channel blocking due to a faster sweeping rate of the gas bubbles generated during

54 Chapter 2. Literature Review operation [112]. Cheng and Scott [88] increased the power density of a DBFC by 3.5 % when moving from a parallel to a serpentine flow field.

Kim et al. [108] used a parallel flow channel in the anode compartment and a serpentine-shaped channel in the cathode. The anode catalyst was a Zr-based AB2-type,

H2 storage alloy and the cathode catalyst was 5 % Pt on C and both were separated by a

Nafion 115 membrane. A current density of 420 mA cm–2 and a power density of 218 mW cm–2 were obtained. The experiment was carried out in a system containing a

–3 –3 solution of 10 wt. % (2.64 mol dm ) NaBH4 in 20 wt. % (5 mol dm ) NaOH fed to the anode at 25 cm3 min–1, and humidified air fed to the cathode at 104 cm3 min–1. They improved the power density of the fuel cell by 56 % by using a gold anticorrosion coating on the stainless steel end plates and on the Ni mesh, which also protected the plates during operation. In a normal MEA, a dead zone can be formed in the anode as a consequence of the generation of hydrogen, which avoids the access of the fuel to the anode [108]. This problem can be alleviated by evacuating the hydrogen bubbles by, for example, leaving a gap between the anode and the membrane through which the hydrogen bubbles can easily escape from the anode. Kim et al. [108] improved the power density of the cell by 27 %, using a corrugated shaped anode separated from the membrane by 2 mm, applying an anticorrosion coating on the cathode channel, and controlling the fuel flow-rate and air humidity. It is thought that humidification of the air helps to remove the accumulation of sodium hydroxide on the surface of the cathode catalyst. Park et al. [113] reported that DBFC performance improves at higher flow rates due to the fact that the hydrogen bubbles are quickly removed from the channels of the fuel flow fields in the anode. The hydrogen bubbles decrease the ionic

55 Chapter 2. Literature Review conductivity of the electrolyte and the effective diffusion coefficients of the reactants, negatively impacting on cell performance, mainly at low flow rates.

Stack configuration

The fuel cell stack consists of:

a) Single cells, where the fluid flow field plates are placed on each side of a MEA

to form the anode and cathode compartments. The plates provide channels for

the reactants and products to flow, and also function as current collectors. In the

stack, the adjacent anode and cathode flow field plates usually function as

bipolar plates. If two or more single cells are placed side by side, a mono-polar

strip stack is formed, where air can be fed to one side of the stack by

spontaneous convection [114].

b) Bipolar electrode cells, formed by two cells by interfacing the anode of one cell

and the cathode of the other one, separated by a space, where fuel and products

flow. The union of two or more bipolar-cells is known as a bipolar cell stack

[114]. The different configurations of fuel cell stacks are shown in Figures

2.11(a) and 2.11(b), in which the fuel is fed in parallel and in series,

respectively.

In order to obtain higher output powers, a large number of cells can be stacked together.

By increasing the number of cells, however, the weight and volume also increase, and this has to be taken into account during cell design. Kim et al. [115] assembled a five- cell stack using a corrugated anode with 5 wt. % Pt on a carbon cloth substrate, a cation-exchange membrane and a gold-coated stainless steel mesh to form an MEA of

72 cm2 effective area. A parallel flow field was used for the anode and a serpentine flow

56 Chapter 2. Literature Review

–3 field for the cathode. A solution of 10 wt. % (2.64 mol dm ) NaBH4 in 20 wt. % (5 mol dm–3) NaOH was fed to the anode and air to the cathode. The authors obtained a power density of 200 mW cm–2 using the stainless steel end plates, which was reduced by

12 % when carbon graphite end plates were used. A force was applied on the endplates to ensure firm contact between the bipolar plates; the force applied was higher on the stainless steel plates than on the carbon graphite plates. The stack weight decreased by a factor of 4.2 when using the graphite endplates.

Fuel/products outlet Fuel/products outlet Bipolar Air Membrane Membrane plates Anode Cathode Anode Cathode

End plate End plate Fuel flow inlet Fuel flow inlet

a) b)

Figure 2. 11 Different configurations of flow arrangements in a bipolar flow cell: a) 4-cell stack with the reactant flow circuit feed in parallel flow circuit and b) a 5-cell stack with the reactant feed in series.

57 Chapter 2. Literature Review

2.1.8. Modelling and Simulation

Mathematical modelling and simulation could play an important role in the improvement of the design and performance of the DBFC. It would accelerate the experimental studies and reduce the cost and timescales associated with laboratory tests.

Verma et al. [116] developed a simplified mathematical steady-state model to predict the fuel cell voltage at a given current density in a O2 DBFC. The model takes into account the equation of cell potential and its losses (reaction 2.24), considering the activation, ohmic and concentration overpotentials. The term due to the activation overpotential was formulated using the reaction mechanisms proposed by Morris et al.

[35] in their half-cell studies on Pt electrodes:

(2.25)

(2.26)

(2.27)

The concentrations of and and the fractional coverages of the adsorbed intermediates were calculated considering that the fuel cell is operated at steady state and isothermal conditions; with fixed rate constants, and varying the concentration of sodium borohydride and sodium hydroxide the activation overpotential in the anode was estimated. The cathode overpotential took into account the diffusion of oxygen across the gas diffusion layer (linearization of Fick’s law), to estimate the surface concentration. This however, ignores the diffusion layer in the catalyst layer, which is the overwhelming source of mass transfer limitations. Ohmic resistances were modelled by estimating an ‘overall’ resistance from experimental data, at different temperatures.

58 Chapter 2. Literature Review

To obtain a reasonable qualitative fit to the experimental data, an additional term representing the concentration polarisation in the anode was required.

– Sanli et al. [117] developed a series of similar models for the BH4 /H2O2 DBFC, again ignoring mass, charge and heat transport. The three models differed only in the manner in which the concentration overpotentials were incorporated and whether the cathode was included in the model. Relationships for the activation overpotential were based on

Tafel laws applied to each electrode and concentration overpotentials were derived by including reactant concentrations in the Tafel expression or by introducing a limiting current density. The ohmic losses were characterised by a constant resistance, estimated experimentally along with the reaction constants and the OCV. The range of current density considered was narrow, up to 0.02 A cm–2 (similar to that considered by Verma et al.), and only when the cathode was explicitly included a reasonable fit to the experimental data was possible.

Shah et al. [118] recently developed a physics-based model for an O2 DBFC. For a known constant applied current the model calculated the cell voltage, considering the borohydride oxidation, borohydride hydrolysis, water reduction and hydrogen oxidation occurring in the anode. The cell voltage was calculated by extracting the activation, ohmic and concentration polarizations (included in the activation polarization through the mass transfer resistances) from the OCV, which can be calculated by assuming the net current between anode and cathode zero. The authors suggested that the hydrogen evolution during borohydride oxidation may be due to the water reduction (Tafel-

Volmer-Heyrovsky mechanism), which was thermodynamically favourable at the anode potentials (-0.828 V vs. NHE). The Butler-Volmer equation was used to approximate

59 Chapter 2. Literature Review the current densities from the borohydride oxidation and the water reduction. The reactant concentrations, the anolyte flow rate, the ionomer volume fractions and the membrane/ionomer properties, were compared to experimental observations from the literature. The performance on a Pt/C anode and the performance on Ni anode were also simulated and compared. The results showed an agreement between the polarization curves from the mathematical model and the experimental results from the literature.

The model, however, could be further developed in order to consider the NaOH conductivity and the membrane water content and temperature of the environment. The

- - role played by intermediates such as BH3OH and the adsorption of OH is very important. Therefore, the developed model should also consider the various borohydride oxidation mechanisms proposed in the literature.

Quantum mechanical models can be used to determine the reaction mechanism on heterogeneous catalysis. The catalyst composition and structure can be related to the activity and selectivity. Density functional theory (DFT) is widely used to analyse the

- reactivity of metal and metal-oxide surfaces. As the adsorption free energy of BH4 ions to the electrode surface was identified as a key indicator of electrocatalyst activity and selectivity, its evaluation with DFT can be used to predict catalyst performance.

Rostamikia et al. [52] applied DFT methods to evaluate the reaction mechanism of borohydride oxidation on Au(111) and Pt(111) crystalline surfaces [119, 120]. The oxidation of borohydride ions was simulated using the first-principles determined elementary rate constants and a microkinetic model. A method based on DFT calculations is developed to approximate the potential activation barriers, which were made potentially dependent using the Butler-Volmer equation. Linear sweep voltammograms were simulated suggesting that species containing B-H bonds are stable

60 Chapter 2. Literature Review surface intermediates at electrode potential where an oxidation current is observed. The presence of BH3 as a stable intermediate was confirmed by surface-enhanced Raman spectroscopy [121].

Rostimikia et al. [52] encourage the use of mathematical modelling to design a more efficient catalyst for the DBFC, by alloying two materials in the appropriate percentage.

A binary metal anode could mix a less active metal for hydrogen generation (Ag or Au) with a more active metal (Ir, Pd, Pt or Ni), maintaining high columbic efficiencies. The aim is to search for bimetallic surfaces that will offer favourable molecular adsorption at a potential of -0.5 V vs. SHE. Cu, Ag, and Au (the Group 11) surfaces give molecular adsorption, whereas metals from Groups 8-11 give dissociative adsorption (adsorption with dissociation into two or more fragments, both or all of which are bound to the surface of the adsorbent) and highly exergonic (it releases free energy). At -0.5 V vs.

SHE, adsorption to Au(111) and Ag(111) surfaces is endergonic, whereas the Cu(111) surface offers molecular and favourable adsorption. Au–Cu binary metals are more stable to oxidation than pure copper, DFT methods were used to predict that a

Au2Cu1(111) surface will oxidize borohydride at an overpotential approximately 0.2 V vs. SHE lower than a pure Au(111) surface. This suggests that AuCu binary electrodes are more active than Au electrodes for borohydride anodic oxidation and may increase the power density and improve the performance of the cell, increasing the current density and decreasing the overpotential [122]. These results agreed with that reported by Yi et al. [58], who deposited Au-Cu nanoparticles at different proportions (Au/C,

Au75Cu25/C, Au67Cu33/C, Au50Cu50/C) on Carbon Vulcan and carried out cyclic voltammetry, chronopotentiometry and chronoamperometry for the borohydride oxidation. The authors conclude that the optimum proportion of Au and Cu for

61 Chapter 2. Literature Review

borohydride oxidation was Au67Cu33 (Au2Cu1), obtaining a current density 46.4 % larger than using pure Au/C. An overpotential of -0.695 V vs. SHE was measured when

Au67Cu33/C was used compared to -0.569 V vs. SHE on Au/C, improving the performance of the DBFC from 20 mW cm-2 using Au/C as the anode material to 51

-2 mW cm using Au67Cu33.

In this work, DFT analysis similar to that carried out by Rostamikia et al. [52] can be used to elucidate the mechanism of reaction of the borohydride oxidation on Pd-Ir(111) surfaces. Further explanation of the computational techniques used will be described in

Chapter 4.

2.1.9. Recycling sodium metaborate product to sodium borohydride reactant

The oxidation of borohydride is an irreversible reaction yielding a complex mixture of borates and metaborates. Different chemical routes exist to recycle these products, which could potentially make the borohydride fuel cell a sustainable process. Kojima et al. [22] investigated the reaction of NaBO2 with MgH2 (reaction 2.28) obtaining yields between 4.8 % and 97 % at pressures of 0.5 MPa (0.5 atm) and 7 MPa (7atm), and a temperature of 550 °C (823 K). The authors found that the yield of NaBH4 obtained from reaction (2.28) increased with temperature and pressure but was independent of the reaction time; obtaining the same yield in 2- and 4-hour experiments. According to

Wu et al. [123] this method has not been developed far enough to offer both high yield and fast reaction rates. The proposed reaction is:

(2.28)

62 Chapter 2. Literature Review

NaBH4 can also be generated from NaBO2 using Mg2Si under high H2 pressure. The yield increased with temperature, obtaining 98 % at 550 °C (823 K) under 7 MPa, though it was the same after 2 hours and after 4 hours of operation [22]:

(2.29)

(2.30)

Coke, as shown in reaction (2.31) or methane (reaction 2.32), can also be used to generate NaBH4 from NaBO2 [22]:

(2.31)

(2.32)

Methane is an inexpensive reducing agent and would generate NaBH4 from NaBO2 in a one step process, reaction (2.32). However, the free energy of this reaction is significantly positive at temperatures between 0 °C (273 K) and 1000 °C (1273 K), which precludes its use for the synthesis of sodium borohydride. For the same reasons, coke or H2 are not used as reducing agents to generate NaBH4 from NaBO2 [123].

Sanli et al. [124] studied the use of hydrogen as a reducing agent to generate NaBH4

–3 and presented a cyclic voltammogram of a solution containing 0.1 mol dm NaBO2 in 1 mol dm–3 NaOH on an Ag gauze electrode, finding a reduction peak at 0.5 V vs. SCE,

- - which the authors attributed to reduction of BO2 to BH4 . Constant potential electrolysis at 0.5 V during 24 and 48 hours at room temperature was employed in order to synthesise borohydride via the following reaction mechanism:

63 Chapter 2. Literature Review

(2.33)

(2.34)

(2.35)

After the electrolysis, a new peak at –0.2 V vs. SCE was found in the cyclic voltammogram. The height of the peak increased with the time of electrolysis and after

3 comparing the CV with that of the same solution with an added 10 dm NaBH4, it was confirmed that the peak was due to the generation of NaBH4 from NaBO2. The reaction was followed by iodometric titration, resulting 9 % conversion of NaBO2 into NaBH4 after 24 h of electrolysis and 17 % conversion after 48 h. Similar experiments were conducted with an Au catalyst, showing the same results but the new peak due to borohydride oxidation was shifted to –0.1 V vs. SCE [124]. Gyenge et al. [125] carried out experiments based on the patent aiming at the electroreduction of borates under both electrocatalytic hydrogenation and direct electroreduction conditions in alkaline media

– showing no measurable amounts of BH4 . The authors also questioned the use of the iodate method to analyse the concentration of borohydride ions because the medium was not acidic enough.

The electro-reduction of borate in aqueous media has been reported in the literature

[126-129], however, Gyenge et al. failed to reproduce the previous attempts and verify that no borohydride was synthesized in these aqueous electrochemical systems.

Thermodynamically, water is not likely to be a good medium for the electrolytic production of sodium borohydride due to the fact that sodium borohydride spontaneously hydrolyses and in alkaline solution the hydrolysis inactivity is due to kinetics. Lower energy is required to generate hydrogen from the reduction of water

64 Chapter 2. Literature Review than it is to reduce metaborates to sodium borohydride (0.83 V compared to 1.24 V, respectively) [123]. Further investigation is required to find a feasible and low cost technique to recycle the borates and metaborates.

2.1.10. Synthesis of sodium borohydride

In order to consider sodium borohydride as a competitive fuel for fuel cells and hydrogen generation, the production cost (US$50 per kg) must be reduced [130].

Research in development of less expensive techniques to generate sodium borohydride is required. Nowadays, sodium hydride (NaH) and trimethyl borate (B(OCH3)3) are used for the industrial production of sodium borohydride [14]:

( ) (2.36)

Reaction (2.36) can achieve 94 % yield and relatively low cost free energy, 167 kJ mol-1

NaBH4. However, the expensive or inefficient production of the raw materials makes the overall sodium borohydride production expensive [123].

The Bayer reaction (2.37) [131] has also been commercially used due to the low cost of the materials required to generate NaBH4. However, due to the high temperature operations (higher than the NaBH4 decomposition (534 10 °C [132]) the Bayer reaction presents some explosion risks.

(2.37)

65 Chapter 2. Literature Review

Investigations have also been focused on the substitution of Na by Mg to reduce the synthesis cost [133], referring to reactions (2.28) and (2.38):

(2.38)

However, further research needs to be carried out in order to achieve high efficiency and fast reaction.

2.2. Sodium borohydride for hydrogen generation

2.2.1. Hydrogen generation from metal hydrides – Sodium borohydride

Hydrides can be divided in binary compounds (MHn, e.g. CaH2) and complex compounds (M(M’H4)n, such as NaAlH4, LiAlH4 or NaBH4). The simple hydrides presume of high thermodynamic reversibility and volumetric energy density, but low gravimetric capacity, whereas the complex hydrides have the reverse properties [134].

The simple hydrides can be hydrolysed to generate hydrogen gas. However, the reaction sub-product at ambient temperature is the metal hydroxide instead of the oxide, meaning that some of the hydrogen atoms will be bonded to form the hydroxide instead of being released as hydrogen gas, decreasing the efficiency [134]. Hydrogen can also be generated from thermal decomposition of metal hydrides, which can be reversed by hydriding the metal under high pressure. Higher energy densities are obtained with this technique than with the hydrolysis, but high temperatures are required. The exothermic nature of the reaction makes it difficult to control, and together with the high temperature operations increase the safety concerns. Among the complex compounds, the borohydrides have received significant attention during the past decades due to their

66 Chapter 2. Literature Review high hydrogen content. The easy handling and high hydrogen content, 10.6 wt. %, together with the capability of producing hydrogen at room temperature, makes sodium borohydride an appropriate hydrogen carrier for portable applications [14].

Peña-Alonso et al. [135] have proposed a mechanism for the catalytic hydrolysis of borohydride ions (reaction 1.2). This mechanism involves two essential kinetic steps: the chemisorption of the borohydride ions on the surface metal (M) and the transfer of

− the electron charge of the adsorbed BH3 to the hydrogen atom of M–H, as shown in

Figure 2.12:

H H H H O O H H H -M -M -M -M - H + BH4 H OH H + H2 -M -M B - -M B H -M B H e - H H e H

Electrode Electrode

Figure 2. 12 Reaction mechanism of the hydrolysis of borohydride ions on a metal surface [135].

Guella et al. [136] proposed a similar mechanism for a Pd catalyst, where they claimed that the activation of the electrode surface towards the hydrolysis is caused by electron- withdrawing and electron-releasing effects of the palladium atoms on the catalyst

− surface via the chemiadsorbed species Pd–BH3 and Pd–H.

The hydrolysis reaction rate increases with temperature and in the presence of acidic substances or boric oxides such as oxalic acid, succinic acid, phosphorus (V) oxide and

67 Chapter 2. Literature Review aluminium chloride. In the presence of oxalic acid and small quantities of water, the reaction could be very violent and the pellets can be heated to incandescence. It has also been observed that the hydrolysis reaction rate is dependent on the concentration of

NaBH4, NaOH and H2O. The concentration of NaBH4 should be as high as possible, taking into account that when the concentration of NaBH4 is high enough, the reaction product (hydrated borates) would precipitate from the solution and block the active sites of the catalyst [137]. Liu et al. [138] reported that for that reason the concentration of

-3 NaBH4 should not exceed 15 wt. % (4 mol dm ), at which borates will still be generated but will not precipitate.

Schlesinger et al. [14] reported that during the hydrolysis of sodium borohydride at ordinary temperatures, the hydrogen generation rate decreases over time due to the

- accumulation of the strongly basic metaborate ion (BO2 ), which increases the pH of the solution. The pH can be gradually decreased, by adding an acid or a catalyst, which will increase the hydrogen evolution rate. On the contrary, and depending on the application it could be convenient to avoid the rapid decomposition of the NaBH4 to produce a constant hydrogen generation for longer time. In that case, the rapid self-hydrolysis of

NaBH4 must be prevented, by dissolving it in slightly basic solution [139].

2.2.1.1. Catalyst materials for the borohydride hydrolysis

Several studies have been carried out in order to find a good homogeneous and/or heterogeneous catalyst for the borohydride hydrolysis [139]. Catalysts based in noble metals, such as Pt, Pd, Ru or Rh, have been widely investigated showing positive

3 -1 -1 3 -1 -1 results. A hydrogen generation rate of 23 dm H2 min gcat (175.6 dm H2 min gmet )

-3 was generated using a Pt/C catalyst and a solution of 10 wt. % (2.64 mol dm ) NaBH4

68 Chapter 2. Literature Review and 5 wt. % (1.25 mol dm-3) of NaOH at 25 oC (298 K) [140, 141]. At the same

3 -1 -1 conditions and concentrations, hydrogen generation rates of 2.7 dm H2 min gcat (270

3 -1 -1 3 -1 -1 3 -1 -1 3 -1 -1 dm min gmetal ), 3 dm H2 min gcat (300 dm min gmetal ) and 5 dm H2 min gcat

3 -1 -1 (41. 73 dm min gmetal ) were obtained using Ru/LiCoO2, Pt/LiCoO2 and Ru-Fe-

Co/AC, respectively [142, 143], indicating that these catalysts are less active towards the hydrolysis than Pt/C. The authors calculated the experimental hydrogen generation rate from the accumulative hydrogen measured versus time; which divided by the amount of catalyst/metal used, considering the catalyst loading, lead to the hydrogen generation rate per gram of catalyst/metal.

3 -1 -1 3 -1 -1 A hydrogen generation rate of 9 dm H2 min gmetal (0.126 dm H2 min gcat ) was obtained using a catalyst formed of Pt and Pd atoms on 150 μm thick carbon nanotubes

-3 (CNT) paper and a solution containing 0.1 wt. % (0.026 mol dm ) NaBH4 in 0.4 wt. %

-3 o 3 (0.1 mol dm ) NaOH at 29 C (302 K), generating around 8 cm of H2 during about 20 minutes [135]. These results are comparable to other noble catalysts that have been

3 reported. Liang et al. [144] generated 3 dm H2 in 15 min with a hydrogen generation

3 −1 −1 rate of 32.3 dm min gRu using 0.3 g of catalyst Ru/C and a solution containing 10

-3 -3 wt. % (2.64 mol dm ) NaBH4 and 5 wt. % (1.25 mol dm ) NaOH.

Ru nanoclusters have been reported to show a high catalytic activity towards borohydride hydrolysis, generating hydrogen constantly as far as the borohydride

3 -1 remains in solution. Hydrogen generation rates varying from 0.1 dm H2 min to more

3 -1 3 -1 -1 than 0.35 dm H2 min , which is equivalent to 96 dm H2 min gmetal , at concentrations of catalyst from 0.2 × 10-3 mol dm-3 to 1.4 × 10-3 mol dm-3, respectively

-3 -3 and using a solution of 0.75 wt. % NaBH4 (150 × 10 mol dm ) and no NaOH [145].

69 Chapter 2. Literature Review

-3 By increasing the pH to 9.6 and using 0.57 wt. % (0.15 mol dm ) NaBH4, hydrogen

3 -1 3 -1 3 generation rates in the range of 0.02 dm H2 min and 0.37 dm H2 min (4 dm min

-1 -3 -3 -3 gcat ) were obtained at catalyst concentrations from 0.08 × 10 mol dm to 1.12 × 10 mol dm-3 Ru(0), respectively at room temperature [146].

Non-noble metals can also be catalytic towards the hydrolysis of borohydride [139].

Chloride salts of manganese, iron, cobalt, nickel and copper are active towards borohydride hydrolysis at room temperature. Although the homogeneous catalysts showed promising results, the metal ions dissolved in the solution can contaminate the product and affect the recycling process [14, 137].

Most of the non-noble metal catalysts for the borohydride hydrolysis are based on cobalt, and usually alloyed with boron. Cobalt and nickel borides are also among the most investigated non-noble metal catalysts, due to high hydrogen yields [137] [147].

-3 -1 -1 o Liu et al. [147] reported 26 dm H2 min gcat generation at 30 C (303 K) with a catalyst composed by Co-B nanoparticles of 10 nm diameter. The very large area of the catalyst was responsible for obtaining higher activity than certain noble metals such as

3 -1 -1 Rh/TiO2. Liang et al. [148] achieved a hydrogen generation rate of 22 dm min g cat using a porous Fe–Co–B catalyst is prepared on Ni foam support and a solution

-3 -3 o containing 15 wt. % (4 mol dm ) NaBH4 and 5 wt. % (1.25 mol dm ) NaOH at 30 C

(303 K). The porous structure of the catalyst facilitated the access of the reactants to the active sites of the catalyst surface. Dai et al. [149] prepared a robust and highly effective Co–W–B amorphous catalyst supported on Ni foam for catalyzing hydrogen

3 -1 generation from alkaline NaBH4 solution. A hydrogen generation rate of 15 dm min

70 Chapter 2. Literature Review

-1 -3 g was obtained using a solution of 20 wt. % (5.3 mol dm ) NaBH4 in 5 wt. % (1.25 mol dm-3) NaOH at 30 oC (303 K).

The H2 generation rate can be increased or decreased varying the concentration of

NaOH in solution [150]. Generally, at low or zero concentration of NaOH, the hydrogen generation rate increases and the hydrogen is dispatched rapidly. However, Kim et al.

[107] used a Co-B catalyst on a porous nickel foam support using different concentrations of NaBH4 and NaOH and did not observe any change in the hydrogen generation rate (around 5.4 dm3 min−1) by changing the concentration of NaOH, except of at 0 wt. % NaOH, when the hydrogen generation rate decreased to 4.75 dm3 min−1.

They attribute this exception to the high generation of hydrogen bubbles in the absence of NaOH, which disturb the borohydride flow to the reactor, making the supply of

NaBH4 solution and lowering the hydrogen generation rate irregular. On the contrary,

Kojima et al. [151] did not find any difference in the hydrogen generation rate when changing the concentration of NaOH from 0 wt. % to 4 wt. % (1 mol dm-3) and used a higher flow rate to feed the borohydride solution into the reactor (200 cm3 min-1). This discrepancy can be attributed to the difference in the fuel flow rate used. At high fuel flow rates, the hydrogen bubbles together with the fuel are impulsed to flow through the catalyst rapidly enough to avoid the disruption in the flow which decreases the contact between the borohydride and the catalyst and causes irregularities in the hydrogen generation rate.

Besides the advantages that a catalyst provides to the system, its degradation is a well- known difficulty. The agglomeration of catalyst particles, surface oxidation or partial dissolution in alkaline NaBH4 solutions are common reasons for catalyst deterioration

71 Chapter 2. Literature Review

[15]. At high hydrogen production rate, the strong generation of hydrogen bubbles in the liquid phase, promotes the collision between catalyst particles that can contribute to damaging the material [137].

2.2.1.2. Accelerators and catalyst precursors

Investigations about the use of acid accelerators to increase the borohydride hydrolysis reaction rate have been carried out to substitute the use of the catalysts [139].

Murugesan et al. [152] investigated the use of several additives to increase the kinetics of the hydrolysis reaction and reported that the highest increase in hydrogen generation rate was obtained by adding 3 mol dm-3 HCl, generating a maximum hydrogen yield of

97% of the theoretical yield, hydrogen per unit weight of material. However, the amount of NaBH4 used for the experiment was not specified by the authors, which makes difficult a further comparison with other publications.

Prosini et al. [153] designed a small hydrogen generation system to fed a fuel cell used in a mobile phone, which delivered 5 cm3 min−1 to match a 1 W power fuel cell. A hydrogen volume of 1.1 dm3 was obtained when the system operated for 2.7 h and using

3 an amount of NaBH4 of 0.5 g and an addition of HCl with 1.6 to 1 volume ratio (cm cm−3). Acetic acid can also be used as an accelerator being a more environmentally friendly and safer alternative to HCl. However, the concentration of acetic acid should double that of hydrochloric acid in order to obtain the same hydrogen generation rate at the same operation conditions and concentration of borohydride [154].

Catalyst precursors can also be used to induce the hydrolysis of borohydride. The catalyst precursor together with NaBH4 is introduced into the reactor. When the water is

72 Chapter 2. Literature Review added, the catalyst precursor is reduced to generate the catalyst for the hydrolysis. Liu et al. [138] reached a maximum hydrogen generation of 1.9 dm3 of hydrogen gas when 1.5 g of H2O mixed with 0.2 g of CoCl2 × 6H2O were added to 1.03 g of NaBH4 according to the following reaction stoichiometry:

(2.40)

In this case CoCl2 catalyses the hydrolysis reaction at the same time as it quickly reacts with NaBH4 to form the catalyst precursor, Co2B, which catalyses the remaining borohydride creating an autocatalytic reaction. However, the most of the hydrogen is rapidly generated in the first step, (reaction 2.40) while the catalyst is liberated from the

3 precursor, i.e. 2370 cm H2 was generated from 1 g NäBH4 through reaction (2.40) with

3 0.1 g CoCl2 × 6H2O, whereas only 19 cm were generated when Co2B was the catalyst

[138]. In another paper, the same authors Liu et al. [147] reached a maximum hydrogen generation rate of 26 dm3 min−1 g−1 using a Co-B catalyst obtained from the chemical

-3 -3 reduction of 0.04 mol dm CoCl2 and a solution of 15 wt. % (0.26 mol dm ) NaBH4 in

5 wt. % NaOH (0.075 mol dm-3) under 30 oC.

Large hydrogen storage capacities have been achieved using either catalyst precursors or accelerators. However, both need replacement after the reaction finishes, increasing the general cost of the hydrogen production. Another inconvenient of the use of accelerators is that once added to the borohydride solution, the reaction is difficult to control.

73 Chapter 2. Literature Review

2.2.2. Engineering aspects of the indirect borohydride fuel cell

The most common design of hydrogen generation systems reported in the literature consists of a tank containing the solution with the fuel, a catalytic reactor and a gas- liquid separator. The fuel solution is pumped through the catalytic reactor, where the hydrolysis reaction takes place. Then the reaction products are conducted to a gas-liquid separator. The hydrogen is collected and ready to be used and the remaining solution is stored in a tank as a by-product to be disposed or recycled [139]. In some cases, the design includes a heat exchanger after the hydrogen generator and a silica drier to reduce the hydrogen steam moisture, as shown in Figure 2.13.

Zhang et al. [1] designed a system consisting of a tubular packed-bed reactor of 2.14 cm internal diameter and 17.8 cm length, and a separate fuel tank containing the borohydride solution (20 wt. % NaBH4 in 3 wt. % NaOH), which was fed to the reactor at a flow rate of 20 g min-1. A commercial BMR06 catalyst supported on a suitable metallic substrate was placed into the packed bed catalyst reactor [155]. The reaction started at room temperature and at constant pressure of 4.83 bar. The temperature increased rapidly at 8 oC s-1 and stabilised at 145 oC (418 K), which remained constant after 265 hours operation, indicating that the catalyst was still active after a long-time operation. The products of the reaction, metaborates in solution and humidified hydrogen gas, were conducted to a heat exchanger to condense the water vapour from the gas stream. The gas coming from the gas-liquid separator passed through a silica dryer bed reactor in order to eliminate the remaining water, as shown in Figure 2.13. A fuel conversion of 92 % was measured after more than 700 hours operation, with periodic start-up and shutdown cycles. High fuel conversion was achieved at a liquid fuel space velocity of 15 h-1 (being the space velocity the relation between the

74 Chapter 2. Literature Review volumetric flow rate and the reactor volume) in a flow reactor operated in a semi- continuous mode, meaning that every hour the cycle is completed 15 times (reactor filled and emptied 15 times) however, lower fuel conversions were obtained at higher fuel space velocities.

Silica drier Heat Fuel Cell exchanger Water inlet Dry H 2 H O Hydrogen 2 generator H2 gas O2

Pump Fuel Excess

Water Gas outlet

H gas Gas-Liquid NaBH 2 4 separator solution tank

Liquid

Figure 2. 13 Arrangement of an indirect borohydride fuel cell (IBFC) system composed by a hydrogen generator from NaBH4, which product feeds a H2/O2 fuel cell [1].

Figure 2.14 shows a further design reported by Zhang et al [105], where the reactor throughput was increased by 200 % via the integration of an immersed heat exchanger into the reactor and increasing the pressure. The outlet stream, which temperature increased due to the exothermic hydrolysis reaction, was passed through the exchanger in order to preheat the inlet fuel before it contacts the catalyst. A fuel conversion of 99

% was obtained at constant temperature profile for different flow rates with similar reactant solution (20 wt. % NaBH4 in 3 wt. % NaOH) [156, 157].

75 Chapter 2. Literature Review

Pre-heated fuel which feeds the reactor Reactor Fuel Cell NaBH 4

H O 2

H /BO2 H2 2 O 2 Fuel Excess

Gas-liquid Separator

By-product

Figure 2. 14 Improved design for hydrogen generation from sodium borohydride, including a heat exchanger inside the reactor for fuel preheating [105].

Kim et al. [158] constructed a hydrogen generation system to feed a fuel cell stack with an output power of 100 W. The system included a lithium battery for power management. The hydrogen generator was similar to the one described in Figure 2.14 but in this case, the pump was powered by the Li battery and the fuel cell stack. They

-3 -3 used a 15 wt. % (4 mol dm ) NaBH4 solution in 5 wt. % (1.25 mol dm ) NaOH and a

3 Co catalyst supported in a gamma type alumina (Al2O3). The fuel was feed at 0.03 dm min-1 which in theory should generate a hydrogen generation rate of 1.070 dm3 min-1 necessary to produce 85 W at the fuel cell stack. The average hydrogen generation rate was 11 % less than expected, 0.946 dm3 min-1 during 60 min, and according to the

76 Chapter 2. Literature Review

authors is due to a decrease in the actual NaBH4 feeding rate produced by the effects of the reactor pressure.

Kim et al. [107] designed two different size reactors to generate hydrogen from sodium borohydride using a Co-B catalyst supported on Ni foam. Different concentrations of borohydride in different concentrations of NaOH were tested at fuel flow rates from

0.01 dm3 min-1 to 0.02 dm3 min-1. As it was expected, the hydrogen generation rate increased with the concentration of NaBH4 and with the catalyst area, but it also increased the generation of sub-product, metaborate [107]. This could be an inconvenient due to the relatively small solubility of the NaBO2, which can precipitate and block the flow channels of the reactor [156]. Moreover, the crystallization of

NaBO2 can disintegrate the catalyst support material [159]. At ambient conditions,

NaBO2 crystallizes in a hydrated form, NaBO2.2H2O or tetrahydrate, NaBO2.4H2O

[160], meaning that less water molecules will be available to react with NaBH4 to generate hydrogen. Investigations have been carried out to find the optimum concentration of NaBH4 to be used in the hydrogen generator reactor.

Shang et al. [161] studied the relationship between the concentration of NaBH4 and the

o precipitation of NaBO2 at different temperatures and concluded that at about 25 C (298

K) the maximum concentration of borohydride to keep the borates from precipitation is between 9 wt. % (2.4 mol dm-3) and 12 wt. % (3.17 mol dm-3). In contrast, Kojima et al.

[162] reported that the concentration of sodium borohydride in water should be kept

-3 below 16 wt. % (4.2 mol dm ) at 25°C, in order to keep NaBO2 in liquid state. In agreement with that, Zhang et al. [156] found that using low concentration of borohydride (15 wt. % ≈ 4 mol dm-3) until the system reaches higher temperatures and

77 Chapter 2. Literature Review then increase the concentration up to 20 wt. % (5.3 mol dm-3), keeps the pressure of the system and prevents borate precipitation. When high concentrations of borohydride are used, the system must be carefully rinsed with water when the reactor cools down, in order to remove the residual borates which may cover the catalyst. By using 25 wt. %

-3 o (6.6 mol dm ) NaBH4 and increasing the temperature to 100 C (373 K), the H2 generation can increase to 5.3 wt. % H2 avoiding the NaBO2 precipitation [160]. A hydrogen generator usually operates between 100 oC (373 K) and 200 oC (473 K) and pressures from 0.4 bar to 14 bar [105]. Thus, the optimum concentration of borohydride to be used for the hydrogen generation depends on the operation temperature.

A number of catalysts based in noble metals, such as Pt, Pd, Ru or Rh, have been widely investigated showing positive results. Peña-Alonso et al. [135] reported the obtaining the highest hydrogen generation rate until 2007, 90 dm3 min-1 g-1; however this is an extrapolated value from an accumulative hydrogen generation of 8 cm3 during

20 min time. The low amount (4.1-4.7 mg) of catalyst (Pt and Pd on carbon nanotubes) used and the standardization of the hydrogen generation rate units (dm3 min-1 g-1), makes an unrealistic assumption for an impractical scaled-up reactor and might lead to not real comparisons with other authors results. Kojima et al. [151] reported a more realistic high scale result, obtaining a constant hydrogen generation rate of nearly 60

3 -1 -3 dm min during more than 8 hours. A solution of 9.1 wt. % (2.4 mol dm ) NaBH4 at a flow rate of 0.12 dm3 min-1 fed the tubular honeycomb monolith (150 × 130 mm) substrate (monolith: 808 g; Pt-LiCoO2: 240 g; Al2O3: 60 g) coated with Pt-LiCoO2 catalyst. The authors report a fuel conversion of 100 % at a constant temperature of 373

K (100oC). For a commercial prototype, the catalysts must be low cost, high catalytic activity and long durability.

78 Chapter 2. Literature Review

2.3. Summary

Borohydrides have received significant attention during the past decades due to their extremely high hydrogen content. Sodium borohydride can be used as a hydrogen carrier, being able to release 90 % of the stoichiometric amount of hydrogen from the hydrolysis reaction [14, 163], or as a fuel for fuel cells, presenting many advantages, such as high theoretical specific energy (17 kW h kg-1) and the availability to release 8e- during oxidation.

The DBFC is a promising power source but many constructional and operational aspects of the system still need to be improved, these include:

a) Finding an appropriate, low-cost, anode catalyst to achieve eight (or close to

eight) electron transfer, while minimizing borohydride hydrolysis.

b) Minimizing reactant crossover or engineering cathode catalysts that are not

active towards borohydride oxidation, perhaps allowing the membrane to be

eliminated altogether.

c) Increasing the current density.

d) Increasing of efficiencies and kinetics of sodium borohydride generation from

sodium metaborates [22, 123].

Besides several investigations on anode materials for borohydride oxidation reported in the literature, the complete oxidation (8 electrons transfer) still remain elusive and further investigations are needed to find a suitable catalytic material. Although it has been claimed that Au catalyses the borohydride oxidation with 8e- transfer and avoids the unwanted hydrolysis reaction, the highest peak power density that has been reported

79 Chapter 2. Literature Review for a DBFC, 680 mW cm–2 at 60 oC, was obtained with Pd/C (2 mg Pd cm-2) anode in a

– BH4 /H2O2 fuel cell with an electrodeposited palladium layer at the anode and a sputtered gold catalyst layer at the cathode [18, 36]. Although the anode material is a very important aspect of the cell, its performance also depends on the cathode materials and the ion exchange membrane [46-48]. As an example, the performance of the cell can be improved from 49.6 mW cm–2 to 77.3 mW cm–2 by the use of Pt/C rather than

-2 Ag/C (2 mgmetal cm ) [49]. The use of different membranes has also been investigated.

The most common membrane used is the Nafion 117 cationic membrane, which possesses excellent mechanical and chemical stability in strongly alkaline media. Other

– types of membrane, such as the Nafion (R)-961, which mitigates the cross-over of BH4 have also exhibited reasonable performance. [90]. The non-commercial membrane polyethylenetetrafluoroethylene (ETFE-g-PSSA) increases the power density and the peak current of the fuel cell but simultaneously decreases the open circuit potential [91,

92]; this is not necessarily a problem if the peak power density increases. Membrane- less DBFCs will be a viable option if an appropriate cathode catalyst not active towards borohydride oxidation can be found, i.e. MnO2 [17].

Another parameter that negatively affects the DBFC performance is the hydrogen generation from the borohydride hydrolysis. As an attempt to minimise the hydrolysis reaction, additives such as TU have been used in the alkaline solution, leading to a variety of conclusions reached by different authors. Most of them agree on the ability of

TU to inhibit borohydride hydrolysis, but they do not agree whether the addition of TU will be beneficial for overall cell performance [19, 95, 96, 98, 99, 164].

The performance of DBFCs can also be improved by varying the operating conditions:

80 Chapter 2. Literature Review

a) Increasing the temperature, which enhances both the current and power

densities. The diffusion coefficients of the reactants and the electrolyte

conductivities increase with increasing temperature. The borohydride

ions will be oxidised at a faster rate, but the rate of hydrogen generation

will also increase [64, 88, 105].

b) Increasing the concentration of borohydride ions, will increase the

current density, but also the rates of hydrogen generation and

borohydride crossover (leading to higher cathode polarizations). An

increase in the concentration of hydroxide ions will reduce the rate of

crossover, but will also increase the solution viscosity, decreasing the

rates of reactant diffusion and increasing ohmic losses through the

electrolytes, particularly at high current densities [48, 88, 106].

c) High fuel/oxidant flow rates, which improves performance by enhancing

reactant transport and reducing channel blocking due to the accumulation

of products [38, 67], but decreases the residence time.

The cell design and construction are also key aspects requiring attention. Fuel flow plate designs with appropriate flow fields for liquid solutions should be optimised for best performance, while avoiding mass-transport issues such as channel blocking. Better results were obtained when using a serpentine flow field rather than the parallel flow field [88, 112], and by using a corrugated anode separated from the membrane by a small gap [108]. The development of mathematical models taking into account the reaction environment, mass, charge and heat transport, together with the complex reactions at the electrodes requires a detailed, physics-based approach. This area of research offers many opportunities since systematic parametric studies could point

81 Chapter 2. Literature Review towards optimal designs (including in relation to materials and geometries) and operating conditions, avoiding expensive and time-consuming trial-and-error experimental procedures.

The needs and requirements when using sodium borohydride as a hydrogen generator for H2/O2 fuel cells are reduced to the increase of the hydrogen generation rate, the durability of an active and low cost catalyst material and the design of a safe/portable hydrogen generator. In order to increase the hydrogen generation rate, parameters such as the operation temperature or the concentration of NaBH4, NaOH and H2O can be optimised. The addition of acidic substances or boric oxides will increase the hydrogen generation rate as well as the presence of an appropriate catalyst for the hydrolysis reaction. Most of the catalysts investigated for hydrogen generation from borohydride hydrolysis are based on cobalt, usually alloyed with boron. However, novel metals such as Pd or Pt also present good performance, being able to generate hydrogen in the order of over ten litres per minute. A simple system design consisting of a reactor with a catalyst bed where sodium borohydride is fed followed by a heat-exchanger and a gas- liquid separator, ideally allows the generation of a hydrogen standalone system [1, 151].

Even simpler systems can be designed including only the reactor connected to the fuel cell making a portable system easy to transport. Investigations need to be carried out in order to find the optimum and affordable catalyst material and conditions and to improve the catalyst durability.

82 Chapter 3: Experimental methodology

Chapter 3: Experimental methodology

3.1. Materials and chemicals

The following materials and chemicals were used as received: sodium borohydride

98+% power (Acros Organics), sodium hydroxide 98% (Fisher BioReagents), sodium sulphite anhydrous (98%, Fisher BioReagents), sodium thiosulphate (>97%, BDH

Laboratory Supplies), Disodium hydrogen orthophosphate anhydrous (Fisher Scientific

UK limited), Nafion ®117 membrane. Triton X-100 (Sigma), zonyl FSO (DuPont), S-

228M anionic fluorosurfactant (Chemguard), sodium dodecyl sulphate, gold (III) chloride, sodium tetrachloroaurate (III) dehydrate (99%, Aldrich), FC4430 (3D). The solutions were prepared using distilled water from a Purite water purification system model Select Fusion 160 BP (0.1 – 5 mS cm-1, electrolytic conductivity). In some cases, still drinking water (pH = 7.7, conductivity = 455 μS cm-1), tap water (Highfield

Southampton, pH = 7.8, conductivity = 630 μS cm-1) or river water (collected from the

River Tawe, Swansea, pH = 7.5 – 8 conductivity = 600 - 800 μS cm-1) or deionised water (pH = 5.5, conductivity ~ 7 μS cm-1) was used.

The catalysts used were: platinised titanium (70 mg cm-2 Pt, layer ≈ 3.5 μm, Magneto

B.V.); palladium iridium alloy (50:50 wt. % loading: 12 mg cm-2 supported on carbon fibres supplied by the ONR [165]); platinum supported on Vulcan carbon printed on a carbon Toray paper loading: 4 mg cm-2 (Johnson Matthey Fuel Cells); Pd deposited on granular carbon (0.5 wt. %, reduced. 4 × 8 Mesh, Specific surface area: 900 – 1100 m2 g-1, Alfa Aesar, Johnson Matthew), planar gold plate, commercial Au/C (10 wt. % Au and 0.5 mg cm-2) from E-Tek.

83 Chapter 3: Experimental methodology

3.2. Equipment

A computer controlled potentiostat/galvanostat PGSTAT30 from Autolab (EcoChemie,

Netherlands) fitted with the General Purpose Electrochemical System GPES 4.9 software was used in all the electrolysis. Varian CP-3800 gas chromatograph.

3.3. Electrochemical cell and methodology

3.3.1. Electrolysis

In order to minimize the hydrogen generation rate during the oxidation of borohydride ions in alkaline solution and explore their effect on the performance of the direct borohydride fuel cell, several anolyte surfactants, including Triton X-100, Zonyl FSO,

S-228M, sodium dodecyl sulphate and FC4430 were evaluated. A table summarizing the properties of the surfactants used is presented in the Appendix I.

The three electrode electrochemical cell shown in Figure 3.1 was used to measure the amount of hydrogen gas generated during the borohydride hydrolysis. The working electrode compartment was gas tight and connected to an inverted 50 cm3 burette, as shown in Figure 3.1. Two types of working electrodes of 3 cm2 geometric area were used: a planar gold electrode and a carbon felt covered with dispersed gold nanoparticles (10 wt. % Au and 0.5 mg cm-2). The planar gold electrode was mechanically polished with P-600 SiC abrasive paper and rinsed with distilled water, followed by an ultrasonic bath cleaning in water before each experiment. A fresh piece of Au/C felt was used for each experiment and each electrode was washed with deionized water and rinsed with isopropanol in order to reduce the surface tension and allow the borohydride solution to penetrate into the pores of the electrode. The electrode was then thoroughly rinsed with water to remove all the alcohol completely.

84 Chapter 3: Experimental methodology

The counter and reference electrodes were Pt mesh (geometrical area: 2 cm2) and

Hg/HgO, respectively.

H2 collection gas burette

RE vs. Hg/HgO (1 mol dm-3 NaOH) CE Solution output

NaBH4 in alkaline solution Pt wire Gas

Pt mesh counter electrode Working electrode (3 cm2)

-3 3 mol dm NaOH WE

Luggin capillary Cation exchange membrane Ground glass flange

Figure 3.1 Assembly used for hydrogen generation measurements during electrolysis of borohydride.

-3 -3 In the anodic compartment, solutions containing 1 mol dm NaBH4 in 3 mol dm

NaOH in the absence and the presence of the surfactants at concentrations from 0.00001 wt. % to 0.1 wt. %, were used. The electrolysis of a blank solution containing only 3

85 Chapter 3: Experimental methodology mol dm-3 NaOH was also carried out for comparison. The counter electrode compartment was filled with a solution of 3 mol dm-3 NaOH and was separated from the working electrode compartment by a protonic Nafion 117 ion exchange membrane.

The electrochemical cell was carefully cleaned with detergent and copiously rinsed with distilled water after each experiment to make sure that there were no residues of the surfactant from the previous experiments. A fresh solution was prepared for each experiment.

Constant current electrolysis within the range of 0.1 A to 1.4 A, corresponding to current densities between 33 mA cm-2 and 533 mA cm-2, and potentiostatic electrolysis between -0.8 V vs. Hg/HgO and 0.4 V vs. Hg/HgO were carried out at room temperature, while the anode potential and the anodic current were monitored, respectively.

In order to quantify the percentage of hydrogen generated during the borohydride oxidation, a gas sample was analysed using the gas chromatograph. The three-electrode electrochemical cell shown in Figure 3.1 was used for this purpose. A gas sampling bag, previously purged with nitrogen, was connected to the air tight anode compartment

-3 -3 (instead of the inverted burette). A solution containing 1 mol dm NaBH4 in 3 mol dm

NaOH and a blank solution containing 3 mol dm-3 NaOH were used. The gas was collected while carrying out electrolysis at constant current (200 mA cm-2) at the planar

Au electrode. From each sampling bag, 50 μl of gas were extracted with a syringe and introduced in the gas chromatograph. The measurements were repeated three times.

86 Chapter 3: Experimental methodology

- 3.3.2. Effects of surfactants on the BH4 oxidation kinetics

In order to calculate effect of the surfactants in the kinetic parameters of borohydride oxidation, cyclic voltammetry at a potential sweep rate of 10 mV s-1 was performed in the absence and in the presence of Triton X-100, SDS, and FC4430 surfactants. A planar gold rotating disk electrode (RDE) and a nanoparticulate gold on carbon (Au/C loaded with 0.5 mg cm-2 Au) attached to a stainless steel RDE were used at a stationary electrode and at rotation rates from 400 rpm to 3600 rpm in solutions containing 0.02

-3 -3 mol dm NaBH4 in 3 mol dm NaOH at room temperature in the absence and in the presence of controlled concentrations of surfactants. The glass RDE cell was assembled as shown in Figure 3.2 and contained a Pt mesh of 2 cm2 geometrical area and Hg/HgO as counter and reference electrodes, respectively. The planar gold RDE (0.125 cm2) was polished with aqueous 0.3 µm particle size alumina and introduced in an ultrasonic bath for 3 minutes followed by thorough rinsing with distilled water.

Cathode Gold compartment RDE electrode SCE electrode

Nafion membrane

Figure 3. 2 Photograph of the experimental arrangement with a gold rotating disk working electrode.

87 Chapter 3: Experimental methodology

The nanoparticulated gold on carbon felt was cut to the size of the stainless steel electrode (0.5 cm2) and clamped with a polyester retaining cap to the RDE assembly, as shown in Figure 3.3. The counter electrode compartment contained 3 mol dm-3 NaOH and a Nafion™ 117 proton membrane was used to separate the counter and working electrode compartments.

Figure 3. 3 Rotating disk electrode assemblies.

3.4. Borohydride oxidation at Pd-Ir/Ti electrode

In order to test a possible more effective anode material for the borohydride oxidation, on which the hydrogen generation rate could be minimised, a Pd-Ir coated on carbon microfibres deposited on a Ti foil was evaluated. The microfibrous carbon anode support was fabricated by a direct charging electrostatic flocking applying technique and provided by Dr. C. Patrissi from the Naval Under Sea Water Fare Centre. The electrode was manufactured by applying a 30-100 kV pulse between a carbon plate containing the carbon microfibers (Mitsubishi Chemical K637212) and a titanium foil

88 Chapter 3: Experimental methodology substrate plate containing a conducting carbon adhesive. The catalyst loading was 12.3 mg cm−2 and consisted of nodules 5–10 mm in height, which showed cauliflower features of 1:1 Pd:Ir composition [165]. Figure 3.4 shows SEM images of the carbon fibers on Ti, A) before and B) after the Pd-Ir catalyst coating.

A B

Figure 3. 4 SEM image of A) carbon fibers deposited by direct charging electrostatic flocking. B) Pd-Ir coated carbon microfiber array [165].

The hydrogen generation rate was measured while applying constant potentials (-0.8 V to 0.4 V vs. Hg/HgO) and measuring the anodic current. The cell shown in Figure 3.1

-3 was used for this purpose, filled with a solution containing 1 mol dm NaBH4 in 3 mol dm-3 NaOH in the anode compartment and with 3 mol dm-3 NaOH in the cathode compartment. Both compartments were separated by a Nafion 117 membrane. Rotating disc experiments were carried out using the Pd-Ir coated on carbon microfibres electrode, using the glass RDE cell shown in Figure 3.2. The electrode was cut with a circular shape of 0.5 cm2 geometric area and clamped with a polyester retaining cap to

-3 the RDE assembly, as shown in Figure 3.3. A solution of 0.02 mol dm NaBH4 in 3 mol dm-3 NaOH was used in the anodic compartment, while a solution containing only

3 mol dm-3 filled the counter electrode compartment. Both, anodic and cathodic compartments were separated by a Nafion 117 membrane.

89 Chapter 3: Experimental methodology

3.5. Gold coated RVC electrodes for borohydride oxidation

3.5.1. PVD sputtering technique

Several reticulated vitreous carbon of 1.0 × 0.5 cm area and 0.5 cm thickness and porosities of 10, 20, 30, 45, 60, 80 and 100 pores per inch (ppi) were gold-coated on both sides during 1, 2 and 3 minutes by physical vapour deposition (PVD) using an

Anatech sputtering system model Hummer® 6.2 15 mA and 8.5 kV in vacuum.

3.5.2. Cyclic voltammetry

The three-electrode electrochemical cell shown in Figure 3.5, with a solution containing

-3 -3 0.02 mol dm NaBH4 in 3 mol dm NaOH, was used.

SCE Reference electrode Pt cathode electrode (0.5 × 0.5 cm) 20 × 10-3 mol dm-3 -3 NaBH4 in 3 mol dm NaOH

RVC working Electrode: 1 × 0.5 × 0.5 cm

Cation exchange Nafion Luggin Magnetic stirring membrane between two capillary follower silicone gaskets

Figure 3. 5 Three-electrode cell.

90 Chapter 3: Experimental methodology

The gold-coated electrodes, prepared as described in the previous sections, were used as the anode electrode. A platinum mesh of 0.5 × 0.5 cm was used as a counter electrode kept in a separate compartment divided by a cation-conducting Nafion® membrane

(DuPont, NF117/H+) to avoid the decomposition of the borohydride ions. A saturated calomel electrode (SCE) was used as a reference positioned close to surface of the working electrodes via a Luggin capillary. The borohydride solutions were freshly prepared before each experiment. Cyclic voltammetry experiments were carried out between -1.0 V and +0.5 V vs. SCE at potential sweep rates of 20 mV s-1, 40 mV s-1, 60 mV s-1, 80 mV s-1, 100 mV s-1, 120 mV s-1, 150 mV s-1 and 200 mV s-1 using an

EcoChemie Autolab (PGSTAT20) computer controlled potentiostat with a General

Purpose Electrochemical Software (GPES) Version 4.5.

3.6. Hydrogen generator design and testing

Figure 3.6 shows a schematic diagram of the two-compartment glass reactor designed for the present studies, which was based on a 1965 patent by Litz et al. [166]. The reactor design consisted of two chambers: an upper chamber holding the catalyst, where the reaction takes place, and a lower chamber where the fuel (sodium borohydride solution) is stored. Both chambers, of 8 cm internal diameter and 10 cm height, were separated by a conical shape glass barrier which ends in a funnel connected to the bottom of the fuel chamber, were coupled together with a polyester clamp and stainless steel screws. The thickness of the glass used to build the reactor was 0.5 cm, able to support up to 10 bar pressure. The funnel base was used to support a nickel mesh to contain the granular catalyst, when it was used.

91 Chapter 3: Experimental methodology

The borohydride solution allocated in the lower chamber was forced through the funnel into the upper body of the reactor, where it was in contact with the catalyst to start the reaction and produce hydrogen. The lower section included a 3-way valve at the bottom to direct the borohydride solution towards the upper chamber through a funnel. This can be achieved by pressurising the bottom chamber to ≈ 2 bar (2 × 10-5 N m-2) with nitrogen gas and using the 3-way valve to allow the solution being transferred into the upper chamber. Two external pressure relief valves with a set value of 3 bar (3 × 10-5 N m-2) were installed in each chamber for safety. A manual control valve was also installed in the fuel chamber to inject nitrogen to pressurize the system. A stainless steel thermocouple, sheath mineral insulated with a pot seal 1.0 mm diameter, connected to a digital thermometer was used to measure the temperature during operation.

In order to dry out the hydrogen gas stream, a drier containing silica desiccant was installed in line with the exit of the hydrogen gas, as shown in Figure 3.6. The silica desiccant was freshly changed before each experiment. The hydrogen gas passed through the flow meter, was mixed with nitrogen gas for safety then vented through the fume cupboard. In theory, when the 3-way valve is open, it allows the solution to be transferred between the fuel and reaction chambers making the system self-regulated

[166]. Hydrogen is constantly extracted at a rate that is set at the flow meter. If the hydrogen generation rate is higher than the hydrogen extraction rate, the reaction chamber will develop higher pressure than the fuel chamber and part of the solution will return to the fuel chamber. As a consequence, the hydrogen generation rate will decrease. If the hydrogen generation rate is lower than the extraction rate, the pressure in the reaction chamber will decrease and the solution will be transferred from the fuel

92 Chapter 3: Experimental methodology chamber to the reaction chamber increasing the hydrogen generation rate and maintaining equal pressure in both chambers.

H2 H2

Pressure relief valve Drier 2

Clamps and top cover H2

Silica desiccant Upper compartment Thermocouple

Gas flow meter Nickel foil support for the catalyst 40 cm

Pressure relief valve 1 Clamps

Ni mesh catalyst support

N2 Port A to feed the Catalyst 2-way valve to pressurise borohydride the system solution Lower compartment Borohydride solution 3-way valve

8 cm

Figure 3. 6 Three-dimensional schematic diagram of the proposed hydrogen generator reactor [166].

93 Chapter 3: Experimental methodology

H2 gas

Silica desiccant Releif valve

Gas flow Upper meter compartment Lower compartment

Releif valve

3-way valve

Figure 3. 7 Photograph of the built hydrogen generator reactor.

A general experimental method was followed for all the catalysts. Before introducing the solution into the glass reactor, the catalyst was weighed or cut and measured and was placed in the nickel mesh catalyst support, in the glass funnel, inside the upper part of the reactor. The upper part of the reactor was covered with a PTFE plate held with a polyester clamp. The whole system was purged with nitrogen gas through the 2-way

94 Chapter 3: Experimental methodology valve in order to decrease the concentration of oxygen inside the reactor chambers and to check any possible gas leakages. After ensuring that the system was leakage free, a solution of sodium borohydride was prepared and added to the lower compartment of the reactor through the port A, indicated in the diagram of Figure 3.6. Figure 3.7 shows a photograph of the built reactor.

The solution was 350 cm3 volume containing 4 mol dm-3 of sodium borohydride and

0.5 mol dm-3 of sodium hydroxide in deionised water, unless otherwise indicated. In some instances the stabilizer, sodium hydroxide, was not used. With the 3-way valve at the bottom of the reactor closed and the relief valve 1 in place, nitrogen gas was introduced through the 2-way valve located on the right hand side of the reactor in order to increase the pressure to approximately 2.0 bar - 2.5 bar. With the lower chamber of the reactor pressurised, the 3-way valve was opened to force the borohydride solution into the upper chamber. As the liquid fuel contacts the catalyst, the generation of hydrogen starts at the upper part of the reactor and therefore the pressure increases. At this point the valve of the gas flow meter can be opened to adjust to the desired flow rate. The experiment continued by adjusting and taking the readings of the volumetric flow rate, the pressure and when available the temperature, at regular intervals of time until the reaction terminates or the flow rate became very low or zero. The N2 adsorption isotherms of the Pd/C catalyst were recorded and analyzed volumetrically using a Micrometrics Gemini 2375 Surface Area Analyzer. Before the N2 adsorption measurement, the sample was outgassed at 200 C for 1 hour. In order to evaluate the

BJH pore distribution, the Halsey formula was used to calculate the thickness of the liquid nitrogen layer [167]. Distributions were determined for both adsorption and desorption parts of the isotherm.

95 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

A Pd-Ir coated microfibrous carbon supported on a titanium plate electrode was evaluated for borohydride oxidation. The hydrogen generated from borohydride hydrolysis was measured during the electrolysis at constant potential in order to analyse the electrodes activity towards the borohydride hydrolysis. To evaluate if the catalyst would be an appropriate anode material for the DBFC, cyclic voltammetry was carried out to analyse the possible oxidation peaks and current densities obtained.

In order to evaluate the effect of the Ir content of the alloy, DFT calculations were used to elucidate the mechanism of reaction of the borohydride oxidation on Pd2Ir1 and

Pd2Ir2 alloys. The competition between borohydride oxidation and hydrogen evolution on the Pd-Ir alloys was also evaluated and compared with that on pure Pd.

4.1. Electrolysis

A material consisting of Pd-Ir coated microfibrous carbon supported on a titanium plate electrode was tested as an electrocatalyst for direct borohydride oxidation. Electrolysis at constant potential between -1 and 0 V vs. Hg/HgO was carried out for one hour, while measuring the accumulative hydrogen gas generated from the borohydride hydrolysis and monitoring the current density. The experiments were performed at two

-3 -3 concentrations of sodium borohydride: 0.5 and 1 mol dm NaBH4, in 3 mol dm NaOH at 23 oC. Figure 4.1 shows the hydrogen generation rate versus potential at the two borohydride concentrations. At the open circuit potential, -0.99 V vs. Hg/HgO, and at zero current density, the hydrogen generation rates measured were 1.3 and 0.65 cm3 min-1 for borohydride concentrations of 0.5 and 1 mol dm-3, respectively. Thus, the

96 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

hydrolysis reaction occurs upon the contact between the sodium borohydride solution and the Pd-Ir electrode surface.

At potentials between -0.8 and -0.4 V vs. Hg/HgO and a concentration of borohydride

-3 of 0.5 mol dm NaBH4, the hydrogen generation rate diminished almost completely and current densities from 100 to 200 mA cm-2 were obtained. At more positive potentials than -0.2 V vs. Hg/HgO, the gas generation rate slightly started to increase, reaching

3 -1 -3 0.96 cm min , at 0 V vs. Hg/HgO. In contrast, in 1 mol dm NaBH4 a lower hydrogen generation rate was measured at the open circuit potential and higher hydrogen generation rate was obtained at all the other measured potentials. The OCV at

-3 concentration of 0.5 mol dm NaBH4 was -0.99 V vs. Hg/HgO, whereas using 1 mol

-3 dm NaBH4 it was -0.98 V vs. Hg/HgO. This slight drop agrees with the finding that the OCV decreases when the concentration of borohydride increased reported by other authors [31, 88]. Larger amount of hydrogen gas is generated at a higher concentration of NaBH4, however, due to the slightly more positive potential it is assumed that part of the hydrogen generated during the hydrolysis was oxidised contributing thus to the

-3 current. The measured current when using 1 mol dm NaBH4 was 5.5 mA compared to

-3 1.6 mA measured using 0.5 mol dm NaBH4 and the potential range corresponds to that at which H2 oxidation might occurs.

At -0.4 V vs. Hg/HgO, the hydrogen generation rate increased from 0.65 to 1.28 cm3 min-1 when the concentration of borohydride was increased from 0.5 to 1 mol dm-3.

This agrees with the suggestion of Liu et al. [31] that at palladium electrodes, higher efficiency (6e-) is obtained using relatively low concentrations of borohydride, whereas for concentrations higher than 1 mol dm-3 the efficiency decreases (4e-). The

97 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

temperature of the solution increased up to 50 oC (323 K), 70 oC (343 K) and 78 oC (351

-3 K) at -0.4, -0.2 and 0 V vs. Hg/HgO, respectively using 1 mol dm (3.7 wt. %) NaBH4.

The increase in temperature, which is attributed to the heat released during the exothermic oxidation and hydrolysis reactions, clearly increased the hydrogen generation rate [63, 67] as can be seen in curve b) of the Figure 4.1. At 0.5 mol dm-3

o (1.9 wt. %) NaBH4 the operation temperature remained constant at about 30 C (303 K) at all the applied potentials.

6 70 C 1

- b)

min

3 5 78 oC 4

o 2 50 C

Hydrogen generation rate /cm rate generation Hydrogen a) o 1 30 C 30 oC

0 -1 -0.8 -0.6 -0.4 -0.2 0 Electrode Potential, E vs. Hg/HgO / V

Figure 4. 1 Hydrogen gas generation rate vs. potential applied at the Pd/Ir coated 2 -3 microfibrous carbon electrode (3 cm ). a) 0.5 mol dm NaBH4 in 3 mol -3 -3 -3 dm NaOH, b) 1 mol dm NaBH4 in 3 mol dm NaOH.

Figure 4.2 shows a comparison of the hydrogen generation rate versus current for 1 mol

-3 -3 -3 -3 dm NaBH4 in 3 mol dm NaOH (curve a) and 0.5 mol dm NaBH4 in 3 mol dm

NaOH (curve b). The figure also compares the results obtained in this work with curves

98 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

c) and d), reported by Liu et al. [31] using a Pd/C electrode with 8 mg cm-2 catalyst

-3 -3 -3 loading and solutions of 0.52 mol dm in 2 mol dm NaOH and 1 mol dm NaBH4 in 2 mol dm-3 NaOH, respectively.

1 -

min 3 3 4

3 c)

2 b) a)

1 Hydrogen /generationratecm Hydrogen

0 0 200 400 600 800 1000

Current density, j / mA cm-2

Figure 4. 2 Gas generation rate vs. current from electrolysis at constant potential: a) Pd/Ir coated microfibrous carbon electrode using a solution of 0.5 mol -3 -3 dm NaBH4 in 3 mol dm NaOH, b) Pd/Ir coated microfibrous carbon -3 -3 electrode and 1 mol dm NaBH4 in 3 mol dm NaOH, c) 10 wt. % Pd/C -3 -3 and 0.52 mol dm NaBH4 in 2 mol dm NaOH. [31], d) 10 wt. % Pd/C -3 -3 and 1.02 mol dm NaBH4 in 2 mol dm NaOH [31].

All the curves have similar trend: hydrogen generated at zero current, followed by a decrease in the hydrogen generation rate between 100 and 400 mA cm-2 and a dramatic increase of the hydrogen generation rate at current densities from 400 mA cm-2 to higher values. Liu et al. obtained a hydrogen generation rate of 4 cm3 min-1 at 100 mA

-2 -3 -3 cm from the solution containing 1 mol dm NaBH4 in 2 mol dm NaOH (curve d). At

99 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

the same current density, a much lower hydrogen generation rate, 0.1 cm3 min-1, was measured in this work using the same concentration of borohydride in 3 mol dm-3

NaOH. At 300 mA cm-2 Liu et al. obtained hydrogen generation rates of around 2.3 and

3 -1 -3 1.6 cm min , for concentrations of NaBH4 of 1 and 0.5 mol dm , respectively; whereas values below 0.5 cm3 min-1 were obtained using the Pd/Ir coated microfibrous carbon electrode at the same current density. The large difference in the hydrogen generation rate obtained by Liu et al. compared to that obtained in this work is attributed to the fact that Liu et al. not only used a different catalyst but also a lower concentration of NaOH. The presence of Ir in the Pd-Ir alloy used in this work and a higher catalyst loading improved the activity of the catalyst, increasing the oxidation of borohydride and decreasing the hydrogen generation.

Figure 4.3 shows the comparison of the hydrogen generation rate obtained on the Pd/Ir coated microfibrous carbon electrode and those using the Au/C and the Au flat plate

-3 electrodes, curves a), b) and c), respectively when a solution of 1 mol dm of NaBH4 in

3 mol dm-3 NaOH was used. The curve d) in the same figure shows the results obtained

-3 -3 by Wang et al. [95] on an Au/C electrode, with 0.7 mol dm NaBH4 in 2 mol dm

NaOH. The figure shows that at low currents, the lowest hydrogen generation rate was obtained when the Pd/Ir electrode was used, i.e. at 300 mA cm-2, the hydrogen generated on the Pd-Ir electrode was 70 % lower than that obtained on the Au/C.

At currents above 2.1 A the difference in the hydrogen generated on the Pd-Ir catalyst and on the Au/C catalyst is minimised and both of them have the same slope. As the slope of the hydrogen generation versus current is related with the Faraday’s law, the number of electrons released during the borohydride oxidation at both Pd-Ir and Au/C

100 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

should be the same. The drawback of using the Pd-Ir coated microfibrous carbon as anode for the DBFC is the hydrogen release at cero current and OCV. The hydrogen bubbles can hinder the start of the borohydride oxidation and decrease the fuel utilization.

-1

min

3 4

3 a)

2 d) b) 1 cm / rate generation Hydrogen c)

0 0 200 400 600 800 Current density, j / mA cm-2

Figure 4. 3 Gas generation rate vs. current obtained during experiments at constant potential, using different anode materials (3 cm2): a) Au flat plate electrode, b) Au/C electrode, c) Pd-Ir coated microfibrous carbon -3 -3 electrode using a solution containing 1 mol dm NaBH4 in 3 mol dm -3 -3 2 NaOH. d) 0.5 mol dm NaBH4 in 2 mol dm NaOH at Au/C (4 cm geometric area) (20 wt. %) [95].

Figure 4.4 shows the number of electrons released during the electrolysis at constant potential according to equation (4.1), proposed by Wang et al. [95] who proposed a

- mechanism of reaction of BH4 oxidation at Au/C by plotting the H2 generation rate versus the current.

( )

101 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

where I is the measured current and G is the current from the oxidation of the hydrogen generated during the hydrolysis of borohydride ions, calculated using the Faraday’s law

- and considering 2 e transfer. For a complete borohydride oxidation, napp = 8, there is no hydrogen generation and G will be zero. According to Figure 4.4 the number of

-3 -3 electrons transferred during the electrolysis of 0.5 mol dm NaBH4 in 3 mol dm

NaOH was almost 8 for potentials between -0.6 and -0.4 V vs. Hg/HgO, whereas at 1

-3 -3 mol dm NaBH4 in 3 mol dm NaOH the number of electrons decreased towards more positive potentials from 6e- at -0.8 V vs. Hg/HgO to less than 2e- at -0.2 V vs. Hg/HgO.

7 a)

app n 6

3 b) 2

1 Apparent number of electrons,of number Apparent 0 -1 -0.8 -0.6 -0.4 -0.2

Electrode potential, E vs. Hg/HgO / V

Figure 4. 4 Apparent number of electrons released during oxidation vs. electrode potential from the electrolysis of solutions containing: a) 0.5 mol dm-3 -3 -3 -3 NaBH4 in 3 mol dm NaOH, b) 1 mol dm NaBH4 in 3 mol dm NaOH at 23 oC using an Pd-Ir electrode.

4.2. Cyclic voltammetry

- Figure 4.5 shows the cyclic voltammogram of the oxidation of BH4 ions carried out at the Pd-Ir coated microfibrous carbon electrode (3 cm2) in a three-electrode

102 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

electrochemical cell. A CV in a blank solution containing 3 mol dm-3 NaOH without

- NaBH4 was also carried out for comparison with the BH4 oxidation on this electrode material.

60 I II

40 III

-2 V 20

/ mA cm mA / j 0

-20

Current density, density, Current -40 -3 0.02 mol dm NaBH4 IV -3 0 mol dm NaBH4 -60

-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 Electrode Potential, E vs. Hg/HgO / V

Figure 4. 5 Cyclic voltammogram at the Pd-Ir coated microfibrous carbon electrode 2 -3 -3 (3 cm ). Solutions containing 0.02 mol dm NaBH4 in 3 mol dm NaOH and 3 mol dm-3 NaOH.

The first peak, I, obtained at -0.95 V vs. Hg/HgO was observed only in the presence of borohydride and is attributed to the oxidation of the hydrogen adsorbed on the electrode surface [69, 168]:

(4.2)

103 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

The second peak II, at -0.71 V vs. Hg/HgO, is surrounded by small shoulders indicating a complex oxidation process of palladium [169]. This peak is attributed to the adsorption of OH- in the initial stage of the palladium oxide formation and partially overlaps the hydrogen oxidation peak [168]:

(4.3)

The small rise III, which emerges from peak II, might be due to the second stage of formation of the palladium (II) oxide layer on the surface of the catalyst [170]:

(4.4)

(4.5)

Peak IV at -0.79 V vs. Hg/HgO can be attributed to the reduction of the Pd(II) oxide during the reverse scan:

(4.6)

In the presence of the borohydride ions, the peak I at -0.95 V vs. Hg/HgO, was noticeably increased by a higher presence of hydrogen formed due to the hydrolysis of borohydride at the open circuit potential (-0.99 V vs. Hg/HgO). In the presence of borohydride, the increase in current density of peaks II and III was probably due to a mixed oxidation of palladium, borohydride ions and possibly intermediate species formed during borohydride oxidation. Peak IV occurring at -0.79 V vs. Hg/HgO on the reverse scan, observed in both borohydride and blank solutions, can be attributed to the reduction of the Pd(II) oxide to Pd(0). On the reverse scan of the borohydride solutions,

104 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

an additional oxidation peal V was observed at -0.17 V vs. Hg/HgO, which can be attributed to the oxidation of intermediate species of the borohydride oxidation on the Ir atoms. Kiran et al. [70] obtained an oxidation peak at -0.43 V vs. Hg/HgO when they

-3 -3 carried out the CV of the 0.1 mol dm NaBH4 in 1 mol dm NaOH using an Ir/C anode.

In summary, the Pd-Ir catalyst presented good activity towards borohydride oxidation, obtaining a number of electrons between 7 and 8 along a wide potential range (from -

0.8 to -0.1 V vs. Hg/HgO). However, the hydrogen generation at the OCV might be a problem for the cell performance. It seems that the presence of Ir increased the catalytic properties of the anode material. In order to corroborate the positive effect of Ir in the

Pd-Ir alloy and to obtain a better understanding of the mechanism of reaction of the borohydride oxidation on Pd-Ir surfaces, density functional theory (DFT) calculations were used.

4.3. Computational methods

DFT calculations were performed to elucidate the mechanism of reaction of borohydride oxidation on Pd-Ir(111) surfaces. The effect of the Ir concentration in the

Pd-Ir alloy was also analysed by studying the borohydride oxidation mechanism in two different alloys, Pd2Ir1(111) and Pd2Ir2(111) surfaces. The optimum reaction pathway was obtained by the investigation of elementary surface reaction energetics and postulated according to the lowest energy path, considering the activation energies for each electron transferred. As a first approximation, the energy of a transferred electron and the ion adsorption energies were made linearly dependent on the electrochemical

105 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

potential and the energy of the adsorbed species was calculated as their free energy

[120, 171, 172].

All calculations were performed using the ab initio total energy and molecular- dynamics Vienna ab initio simulation program (VASP) developed at the Institute for

Material Physics at the University of Vienna [173-175]. VASP is a computer program for atomic scale material modelling, e.g. electronic structure calculations and quantum- mechanical molecular dynamics, from first principles [176]. VASP computes an approximate solution to the many-body Schrödinger equation (4.7), within DFT, by solving the Kohn-Sham equations [177].

̂ ( ) ( ) (4.7)

Where is the energy of the state , which is the wave function of the quantum system for the N electronic eigenstates, being N the number of electrons in the molecule, and related to the electronic density by ∫ ( ⃗). In quantum mechanics is the rough equivalent in classical mechanics to the knowledge of the position and momentum of a particle and 2 is the probability per unit volume of finding the particle in the volume element dx,dy,dz. The particle is normally confined in an imaginary box. ̂ is the Hamiltonian operator, which characterizes the total energy of any given wave function. By analogy with classical mechanics, the Hamiltonian is commonly expressed as the sum of operators corresponding to the kinetic ( ) and potential ( ) energies of a system, in the form:

106 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

(4.8)

The determination of the energy for N electrons would require a large number of computational operations impossible to solve in a reasonable period of time. To simplify these equations and reduce the calculation time, the Hartree-Fock theory [178], which allows solving N coupled one-electron equations instead of solving one

Schrödinger equation for N electrons, can be used. The one-electron Schrödinger equation can be written as follows:

̂ ( ) ( ) (4.9)

where ψj is the wave function of the electronic state j, r is the spatial coordinate relative

th to the j electron, and is the Kohn-Sham eigenvalue for one-electron j.

In order to solve equation (4.9), Kohn and Sham developed a series of equations to calculate the wave function, the electronic density and the ground state energy of a particle. To calculate the total energy for the N electrons, two theorems given by

Hohemberg and Kohn should be considered [179]:

 Only one ground-state electron density corresponds to one Hamiltonian and

thus, it is possible to define the ground-state energy as a functional of n(r): =

[n].

 [n] is minimal when n(r) is the ground-state density, among all possible

electron densities. The minimization of [n] with respect to the density gives

the ground state energy of the system.

107 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

It is not possible to calculate without a previous knowledge of the wave function.

However, there is no simple function or relationship between and n(r) that does not involve the wave function. In order to calculate the energy at the ground state, it is necessary to determine the set of wave functions that minimize the Khon-Sham energy functional. The steps that VASP follows to calculate the energy at the ground state for a

- molecule or ion, such as BH4 , are detailed below:

Step 1- Calculation of the wave function:

The Kohn–Sham equation (4.10) considers the Schrödinger equation in a fictitious non- interacting system:

[ ( )] ( ) ( ) (4.10)

where h is the Plank constant, m is the particle mass, is the vector of partial derivatives (∂f / ∂x1, ∂f / ∂x2…, ∂f / ∂xn), and Vs(r) the external potential. The first term is referred to the kinetic energy and the second is due to the potential energy. Both terms form the Hamiltonian operator.

Kohn-Sham considered that for any system of N interacting electrons in a given external potential ( ), there is a virtual non-interacting system, of the same number of electrons, with exactly the same density as the interacting system. In that case, the virtual non-interactive system must be subjected to another external potential Vs(r), which includes the contribution from the electron-electron interactions. Thus, the non- interactive system potential Vs(r) can then be corrected by considering the electron- electron interactions and the external potential for the interacting electron system,

( ):

108 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

( ) ( ) ( ) ( [ ]) (4.11)

Substituting equations (4.11) in equation (4.10), leads to equation (4.12):

[ ( ) ( ) ( )] ( ) ( ) (4.12)

where ( [ ]) is the exchange-correlation potential, which is the functional derivative of the exchange correlation energy and it is also a functional of the electron density n( ):

( [ ]) [ ] ( ) (4.13)

The exchange-correlation energy considers the electronic correlation, which is the interaction between electrons in the electronic structure of a quantum system, and the exchange interaction, which is a quantum mechanical effect between identical particles.

( ) is the Hartee potential of the electrons, which is referred to the Coulombic potential, the potential due to other electrons in which electron j moves and can be defined as:

( ) ( ) ∫ (4.14) | |

where e is the elementary charge of the electron, d is the length of the box where the particle is, r’ is the position of the other electron different to j. Equation (4.12) can be developed as follows using equations (4.13) and (4.14):

109 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

( ) [ ( ) ∫ [ ] ( )] ( ) ( ) (4.15) | |

The wave function can be calculated by solving equation (4.15), starting from a trial potential or electronic density and iterate to self-consistency. In this case, a first trial electronic density n1(r) was guessed by VASP so the wave function can be calculated as follows.

Step 2- From the wave function to the electronic density

If the wave function is known, the electronic density can be calculated by:

( ⃗) ∑ | ( ⃗)| (4.16)

Once the electronic density and the wave function are calculated, the energy at the ground state can also be found.

Step 3- Calculation of the total energy at the ground state

The total energy at the ground state can be computed as a functional of the electronic density according to the Kohn–Sham equation.

[ ] [ ] ∫ ( ) [ ] [ ] (4.17)

where [ ] is the Hartree term describing the electron-electron Coulomb repulsion:

( ) ( ) [ ] ∫ ( ) ( ) ∫ (4.18) | |

110 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

and the kinetic energy term, [ ], can be defined by:

[ ] ∑ ( ) ( ) (4.19)

The exchange-correlation energy was calculated by using equation (4.19), that uses the generalized gradient approximation (GGA) based on that of Perdew-Wang [180].

Considering that in a real non-homogeneous electron system, the electron density varies with space; the generalized gradient approximation suggests that the energy is a functional of both the density and its gradient expansion. This theory supports that two electrons cannot be in the same energy level with the same spin.

[ ] ∫ ( ( ⃗) ( ⃗)) (4.20)

If equations (4.18), (4.19) and (4.20) are substituted in equation (4.17), the Kohn-Sham total-energy functional for a set of doubly occupied electronic states can be written as [181]:

[{ }]

( ) ( ) ∑ ∫ ( ) ∫ ( ) ( ) ∫ [ ( )] | |

({ }) (4.21)

where ({ }) is the coulombic energy associated with interactions among the nuclei (or ions) at positions { }. Equation (4.21) considers the kinetic energy and the

111 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

energy due to: electron-electron interaction, nuclei-nuclei interaction, nuclei-molecule interaction and the exchange in correlation.

In order to calculate the three steps previously described, VASP uses a self-consistent method, in which some values are previously guessed and then confirmed or recalculated, as follows. Figure (4.6) shows a flow chart summarising the computational procedure for the calculation of the total energy at the ground state. The self- consistency starts by a trial value of electronic density ( ), which VASP guesses from the position of the molecule and the number of electrons previously introduced in the programme. With that first trial value ( ) and the equation (4.15) the wave function ( ) is calculated. With the calculated value of wave function ( ) and equation (4.16) the electronic density ( ) can be calculated to check if the guessed charge density and the calculated ones are the same. With both electronic densities ( ) and ( ), the total energy at the ground state and can be calculated using equation (4.21). If = , the derivative of the energy with the position will be calculated. VASP slightly changes the nuclei position and calculates the energy for the new position of the molecule, and that would be the derivative of the energy versus position. If that derivative is close to zero or lower than a small predetermined value (0.05 eV), then, the energy at the ground state was found. The ground state energy of the system of N electrons with the ions in positions { }, can be calculated by minimising the Kohn-Sham energy functional (d /dr ≈ 0).

112 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

Trial electronic density value n1( ⃗), from the molecule position and number of electrons introduced in VASP

Use equation (4.15) to find ( )

With ( ) and equation (4.16)

calculates n2( ⃗)

Plug n1( ⃗) and n2( ⃗) into equation (4.21) and check if: NO ( ⃗) ( ⃗)

YES

NO Check d /dr ≈ 0

YES

Calculation done:

= Energy at the ground state

Figure 4. 6 Flow chart describing the computational procedure for the calculation of the total energy at the ground state.

The energy at the ground state can be calculated for all the molecules involved in the borohydride oxidation mechanism. The energies at the ground state of two molecules forming an elementary reaction, which involves one electron, correspond to the energies of the initial and final states of that reaction. In the case of the borohydride oxidation, 8 elementary reactions lead to the overall borohydride oxidation reaction.

113 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

The energy at the ground state of all the molecules possibly involved in the borohydride oxidation can be calculated, as previously described, in order to elucidate what molecules will be intermediates of the reaction, those with the lowest energy values.

The activation energies of each elementary reaction, between the initial state and the final state, can be calculated to estimate what mechanism is more favourable, that with the lowest activation barriers. The following sections describe these calculations and their use for the elucidation of the mechanism of reaction of the borohydride oxidation.

4.3.1. Pd-Ir(111) crystalline structures

The Pd-Ir bulk and surfaces were built and visualized with Material Studio 6.1. To optimize the structure in VASP, a 3 × 3 × 1 Monkhorst-Pack grid [182] was used, followed by a 5 × 5 × 1 grid single-point calculation to give the total energy. The

Monkhorst-Pack grid technique automatically generates a grid, which number of parts varies with the K-points, and it describes how the calculation is divided over reciprocal space. A higher number of K-points make the calculation more accurate but it is more time consuming. For that reason, convergence tests were carried out to find a balance between the calculation time and the accuracy.

The Pd2Ir1 and Pd2Ir2 bimetallic surfaces were modelled using a 4 layer slab and a 3 × 3 surface cell and a 4 layer slab with a 4 × 4 surface cell, respectively; with the two bottom layers constrained to fcc lattice position and the two top layers relaxed. A vacuum of 13 Å was inserted between the periodic slab representations of the electrode surface to evaluate species interactions with the electrode. Anderson and Kang [171] and Nørskov and et al. [172] have used similar methods, with various perturbations, to

114 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

study a number of electrocatalysis reactions. The borohydride molecule and all the possible borohydride oxidation intermediates were optimized separately on both Pd2Ir1 and Pd2Ir2 surfaces.

In order to calculate the wave function, the projected augmented wave method [183,

184] and a plane wave basis set were employed using an energy cut off of 450 eV. The energy cut off is the maximum allowed kinetic energy of the plane waves involved in the calculation of the wave function. It determines the number of plane waves in the basis set. More accurate results are obtained using high cut-off energy, but the number of Fourier coefficients to be calculated becomes bigger and stacked in the computer, enlarging the time required to solve the calculations.

Once VASP provided the energy of the adsorbed species, the adsorption energies of

- BH4 and its oxidation intermediates were calculated at the optimal adsorption configurations and were corrected for spurious slab-to-slab dipole interactions along the surface normal direction. For comparison with experimentally obtained data, zero-point vibrational energies (ZPVEs) are required to convert total electronic energies obtained from ab initio quantum mechanical studies into 0 K enthalpies total energy of adsorbates. This conversion can be calculated using the harmonic vibrational modes, which considers the translational and rotational motion of the molecule as a hole and can be calculated in VASP. After that, a correction of the temperature must be done.

The free energies of adsorbed species were determined by adding ZPVE and subtracting vibrational entropy (SvibT, at 298 K) from the electronic energy (energy given by VASP,

) [119]:

115 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

* GA = + ZPVE – SvibT (4.22)

- - The free energies of aqueous phase molecules and ions such as BH4 , BO2 , H2O and

OH- were previously reported by Rostamikia et al. [120], who calculated those values by adding ZPVE, pressure volume (PV) and entropic terms to the isolated gas phase molecule energy ( ). Experimental solvation free energies and concentration adjustments were included to consider solution conditions assuming an ideal solution.

4.3.2. Activation energy

The activation energy for each step of the borohydride oxidation reaction on Pd-Ir surfaces were calculated using the lowest energy adsorbed states, presuming each step follows the minimum energy path [119, 121]. The transition states of elementary reactions on the surfaces were isolated with the climbing image nudged elastic band method (CI-NEB) [185], which finds saddle points and minimum energy paths between the known reactants and products. A number of equally spaced images (usually 4 images and occasionally 8) between the initial and final states were optimized. The lowest energy possible for each image was found and the highest energy image corresponds to the transition state, which gives the activation energy of the reaction.

The non-potential dependent activation barriers were extrapolated, using a Butler-

Volmer formalism and an approach recently designed in the Janik group [121], in order to make them potential dependent. A symmetry factor is used to linearly relate the activation barriers to the equivalent electrochemical reaction [119]:

( ) ( ) (4.23)

116 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

where is the activation barrier for the chemical step, is the symmetry factor which will be considered 0.5, as it is the average between 0.3 and 0.7, which are the values in most electrochemical systems [186]. Vo is the equilibrium potential for the oxidation of surface hydrogen.

4.4. DFT results and discussion

4.4.1. Borohydride ion adsorption over Pd2-Ir1(111) and Pd2-Ir2(111)

Figures 4.7(a) and 4.7(b) show the optimum configuration for the molecular adsorption of borohydride ions on Pd2Ir1(111) and Pd2Ir2(111) surfaces, respectively.

a) b)

- Figure 4. 7 BH4 molecule adsorbed on: a) Pd2-Ir1(111) and b) Pd2-Ir2(111). The dark blue atoms correspond to Ir, the light blue slightly larger atoms correspond to Pd, the white atoms are H atoms and the pink atoms are B.

The BH4 species is adsorbed on the Pd2-Ir1(111) slab with the boron atom above the Pd-

Pd bridge and two H atoms on atop sites, as shown in Figure 4.7(a). This was the only

* stable configuration found for a molecularly adsorbed borohydride specie (BH4 ) on

* * * Pd2-Ir1(111). The activation energy for the decomposition of BH4 to BH3 + H is negligible, dissociative adsorption occurs spontaneously and is thermodynamically

* * favourable (ΔGads < 0), producing BH3 + H . The adsorption reaction can be written as:

117 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

( ) ( ) (4.24)

where the * represents the Pd-Ir bare surface. As shown in Figure 4.7(b), the borohydride ion is adsorbed on the Pd2Ir2(111) surface with the boron atom above the hollow site, two hydrogen atoms on the Pd atoms atop sites and one hydrogen atom on the Ir atom atop site. The adsorption reaction of the borohydride ion on the Pd2Ir2(111) surface corresponds to reaction (4.25):

( ) (4.25)

The free energy of adsorption of borohydride ions on the Pd2-Ir1(111) and Pd2-Ir2(111) surfaces can be calculated using equations (4.26) and (4.27):

(4.26)

(4.27)

where is the free energy of the borohydride ions in aqueous solution, is the

free energy of the bare surface, is the free energy of the dissociatively

adsorbed borohydride species, is the free energy of the adsorbed molecule, -e is the charge of an electron and U is the absolute (vacuum reference) electrode potential.

The absolute potential, U, can be referred to a standard hydrogen electrode using the equation:

(4.28)

118 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

where USHE is the potential versus the standard hydrogen electrode (SHE). Note that

USHE is what we named E vs. SHE in the experimental sections. According to the literature [187], the has an average value of 4.6 V referenced to the absolute vacuum. That value is consistent with that calculated by Rostamikia et al. [119] using

DFT methods. Figure 4.8 plots the adsorption energy of borohydride ions over Pd2-

Ir1(111) and Pd2-Ir2(111) versus potential. One adsorbate molecule per nine surface atoms was considered on the Pd2Ir1 surface and one adsorbate molecule per sixteen

- surface atoms on the Pd2Ir2 surface. The adsorption free energy of BH4 over the

Pd2Ir1(111) and Pd2Ir2(111) surfaces were -1.12 eV and -0.89 eV, respectively at -0.5 V vs. SHE (-0.64 V vs. Hg/HgO). That suggests that the presence of Ir in the alloy increases the adsorption energy, making it less exergonic (releases less energy). These values were unexpectedly lower than that reported on pure Pd and pure Ir, -2.16 eV and

-2.01 eV, respectively at -0.5 V vs. SHE [122].

0.0

-0.5

-1.0

/ eV / -1.5

Ads

-2.0

-2.5 Pd2Ir1

Pd2Ir2 -3.0 -1.0 -0.5 0.0 0.5 1.0 Electrode Potential, E vs. Hg/HgO / V

- Figure 4. 8 Free energy of adsorption of BH4 aq over Pd2Ir2(111) and Pd2Ir1(111) - -3 using vacuum slab model (T = 298 K, [BH4 aq] = 0.03 mol dm ).

119 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

The adsorption of borohydride over the Pd-Ir(111) surfaces is stronger (lower in energy) than that on Au(111) and weaker (higher in energy) than on Pt(111). According to the literature [119] the adsorption of borohydride ions at Au(111) starts being favourable (∆G < 0) above 0.05 V vs. SHE, whereas at Pd-Ir(111) the adsorption of borohydride is favourable at any potential between -1 and 1 V vs. SHE. That suggests that Pd-Ir(111) is active towards the oxidation reaction at more negative potentials than

Au(111). The borohydride adsorption energy on Pt(111) was -1.85 eV, at -0.5 V vs.

* * SHE, which causes the molecule dissociation to BH + 3H [120]. At Pd2Ir1(111), the borohydride adsorption is less negative than on Pt(111) and the molecule is adsorbed

* * and spontaneously dehydrated to BH3 + H , whereas at Pd2Ir2 the borohydride molecule was stable and optimized as BH4. It seems that the presence of Ir makes the borohydride adsorption energy more positive and less likely to be dissociative [120,

122].

- 4.4.2. Elementary surface energies of BH4 oxidation at Pd2Ir1(111) and Pd2Ir2(111) surfaces

After the borohydride adsorption energy was calculated, the free energies of all the other electro-oxidation steps can be calculated using the following equation:

(4.29)

where R* is the reactant adsorbed on the surface, P* is the adsorbed oxidized product and n is the number of electrons exchanged. The free energy of the reaction (4.29) can be calculated as:

120 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

(4.30)

The reaction energy diagrams for the borohydride oxidation over Pd2-Ir1(111) and Pd2-

Ir2(111) surfaces were evaluated considering the relative free energies of all adsorbed species with respect to the free energy of the initial aqueous borohydride ion. The relative energy of each species was calculated by combining equations (4.27) or (4.28) and (4.30), leading to equation (4.31) [120].

( ) ( ) ( )

( ) (4.31)

where 0 < x < 2 and 0 < y < 4. The potential dependent activation barriers were also calculated for each oxidation step. As an example, Figure 4.9 shows the initial state (a),

* * transition state (b) and final state (c) of the dehydrogenation reaction from BH3 to BH2

* -1 + H on Pd2Ir1. The activation barrier for this reaction was 0.09 eV (kJ mol ) at -0.64 V vs. Hg/HgO (-0.5 V vs. SHE). The dissociating B-H bond stretches from 1.2 Å in the initial state, to 1.35 Å at the transition state. The initial, transition and final state of the same reaction on Pd2Ir2 is illustrated in Figure 4.10 (a), (b) and (c), respectively. In this case the activation barrier was 0.36 eV (kJ mol-1), which indicates that this reaction step is more difficult to perform on Pd2Ir2 than on Pd2Ir1.

The activation energies for each oxidation step were calculated together with the relative energy diagrams in order to evaluate the most likely reaction mechanism for the borohydride oxidation.

121 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

a) b) c)

* * * Figure 4. 9 BH3 dehydrogenation to BH2 +H over Pd2Ir1(111) surface: a) initial state, b) transition state, c) final state. The dark blue atoms correspond to Ir, the light blue slightly larger atoms correspond to Pd, the white atoms are H atoms and the pink atoms are B.

a) b) c)

* * * Figure 4. 10 BH3 dehydrogenation to BH2 +H over Pd2Ir2(111) surface: a) initial state, b) transition state, c) final state. The dark blue atoms correspond to Ir, the light blue slightly larger atoms correspond to Pd, the white atoms are H atoms and the pink atoms are B.

Figure 4.11 shows the relative energies of all the possible intermediates of reaction of the borohydride oxidation on Pd2Ir1(111) at -0.64 V vs. Hg/HgO (-0.5 V vs. SHE). The most favourable mechanism is highlighted in the diagram with a red straight line. The broken and dashed lines represents the other possible mechanisms calculated with DFT, the most favourable mechanism was determined assuming that the reaction follows the lowest activation energies and the intermediate molecules with the lowest relative energies are involved in the reaction.

122 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

Oxidation reaction coordinate 0.5

BH4 0.88 0 0.28 -0.5 0.09 0.96 BH3OH 0.56 -1 BH3 0.71 0.19 -1.5 0.17 BH2OH

BH2

SHE / eV / SHE

vs. -2 0.63 B 1.29 -2.5 BH 1.75

0.5 0.5 V 1.28

- 1.04 at -3 0.7 0.84 0.15 BO -3.5

Relative Energy for borohydride oxidation borohydride for Energy Relative -4 BHOH BOH BOOH

-4.5 B(OH)2 BO

B(OH)3 -5

- -3 - -3 Figure 4. 11 Reaction energetics of the borohydride oxidation over Pd2Ir1(111) at -0.5 V vs. SHE, [BH4 ] = 0.02 mol dm , [OH ] = 2 mol dm , T = 298 K. The red line represents the most favourable mechanism of borohydride oxidation according to the relative energies and activation barriers calculated from DFT. The broken and dashed lines represent other possible mechanisms of borohydride oxidation analysed. Numbers on the plot label the activation barrier of each elementary step.

123 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

According to the red line, the 8e- transfer seems to be favourable since the reaction energy is downhill, as each intermediate species has lower relative energy than the previous one.

In general, the B-H breaking was more favourable than the B-OH binding and it dominates the oxidation process until the formation of BH*. The first reaction step,

* * * * from BH4 to BH3 , is barrier-less, the energy of BH3 is lower than that of BH4 . The

* * * second step considers two possibilities: 1) BH3 might react with OH to form BH3OH

* * or 2) it might dehydrogenate to form BH2 + H . The latter is more favourable since the

* reaction product BH2 is more electronegative (lower relative energy) and thus more

* stable than BH3OH , and the activation barrier of the dehydrogenation is 0.09 eV,

* compared with 0.88 eV of the BH3OH formation. The same reasoning was followed in all the steps to decide which mechanism of borohydride oxidation is the most likely to

* * * occur. Then, the dehydrogenation from BH2 to BH +H takes place.

The next step involves the OH* molecule reacting with BH* to form BHOH*, followed by a dehydrogenation process with a low energy barrier forming BOH* + H*. From that point forward, the activation barriers are high enough to conclude that B-OH bonds do not form at a substantial rate at room temperature and -0.5 V vs. SHE. Several transition state searches considering different reaction trajectories were performed to search for paths with lower activation barriers which allow the reaction to reach 8e- release; however, the lowest activation energies found are reported herein and suggest that B-OH formation will be challenging at these conditions.

124 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

There are two possible interpretations for these results:

1) The reactions of intermediate products with hydroxyl ions limit the overall

oxidation reaction. The barriers from BOH to further oxidation will limit the

overall oxidation rate and are possibly too high for the reaction to go forward,

leaving BOH as the final oxidation product. Being a stable species, BOH, might

cover the catalyst surface and poison it, reducing the catalyst activity towards

the borohydride oxidation. This results are understandable considering that the

- number of electrons released per BH4 on pure Pd and pure Ir have been

reported to be less than 8, and thus in a Pd-Ir alloy an 8 electron transfer was not

expected. The number of electrons released according to this mechanism of

oxidation on Pd2Ir1 would be 5.

2) The barriers for B-OH forming reactions from the DFT approach used are

generally too high due to the lack of surface species solvation in the surface

model. The complete oxidation reaction is favourable since all the intermediate

products are lower in energy than the previous one, however, the high barriers

hinder the reaction to reach the final product and release 8e-. Solvation would

lower the barrier in two ways, water molecules could more strongly solvate the

* transition state than initial state or the OH could be shuttled through a H2O

* molecule (H from H2O transfers to OH while a B-OH bond is formed with the

O of water). Nie et al. [177] studied the kinetics of elementary steps in CO2

electroreduction, finding that for all O-H bond formation reactions, including

those involving C-OH bond dissociation, the activation barrier is higher than

that of H shuttling reaction. Computational models constructed by explicitly

incorporating the role of water in a computationally tractable manner were

125 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

utilized. Activation barriers for all the elementary steps in the reaction path were

examined for both the water-solvated and H-shuttling. The presence of two H2O

molecules next to the CO2 and intermediate species led to lower C-OH bond

formation barriers, whereas it did not affect the reaction steps involving H

breaking. It might be possible that lower activation barriers can be found when

considering water molecules close to the borohydride oxidation intermediate

steps, leading to easier B-OH bond formation and being able to reach the final

oxidation product, metaborate, and releasing 8e-.

Figure 4.12 shows all the possible intermediates and activation barriers for the oxidation of borohydride ions on Pd2Ir2(111). The preferable mechanism of reaction is highlighted with red lines. The reaction starts with borohydride adsorption, which was

* * * * optimized as a stable BH4 . The transition from BH4 to BH3 + H was barrier-less,

* meaning that BH3 formation will spontaneously occur. The next oxidation step,

* * oxidation of BH3 had only one possible product, BH2 , since the intermediate molecule

* BH3OH was unstable and dissociated during the optimization process. An activation

* * * barrier of 0.36 eV allows the BH3 molecule to form BH2 + H , followed by a dehydrogenation process to form BH*, which would tend to overcome 0.45 eV to form a more stable product (BHOH*), rather than overcome a slightly higher barrier 0.49 eV to form a less stable B*. From BHOH* a low barrier of 0.19 eV led to the formation of

* * BOH + H . Similar to the mechanism of borohydride oxidation on Pd2Ir1(111), from this point the activation barriers to find the final product (metaborate) becomes higher.

* * However, BOH will probably react with OH overcoming 0.55 eV to form B(OH)2. It is possible that at this point the reaction does not go forward due to the high activation

* * * barrier between B(OH)2 and B(OH)3 and the final product will be B(OH)2 .

126 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

Oxidation reaction coordinate 0 0.71 BH4 0.36

-0.5 0.65

-1 BH BH 0.18 0.49 3 2 0.7 1.32 0.45 1.22 -1.5 1.4 BH 0.23 -2 0.55 0.88 SHE / eV / SHE B

0.19 vs. BH2OH -2.5 0.62

0.5 0.5 V BOOH - BHOH BO

at BOH -3

B(OH)2

-3.5 Relative Energy for borohydride oxidation borohydride for Energy Relative B(OH)3 BO2 -4

- -3 - Figure 4. 12 Reaction energetics of the borohydride oxidation over Pd2Ir2(111) at -0.5 V vs. SHE, [BH4 ] = 0.02 mol dm , [OH ] = 2 mol dm-3, T = 298 K. The red line represents the most favourable mechanism of borohydride oxidation according to the relative energies and activation barriers calculated from DFT. The broken and dashed lines represent other possible mechanisms of borohydride oxidation analysed. Numbers on the plot label the activation barrier of each elementary step.

127 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

* * If the activation barrier of the B(OH)2 + OH reaction was lower, a final product

B(OH)3 might react in the solution phase to produce the hydrated borate anion, hydrated form of BO2 [121]:

( ) ( ) ( ) ( ) ( ) (4.32)

Then, the reaction mechanism would be finished with the release of 8e-. However, further investigations are required to analyse the effect of the solvating conditions, which might decrease the activation barriers in the steps involving OH. Experimental techniques, such as Raman spectroscopy would also help to demonstrate if the final product is B(OH)2, as it would be adsorbed on the electrode surface.

Comparing the two diagrams shown in Figures 4.11 and 4.12, it can be seen that the reaction mechanism is more favourable on Pd2Ir2 than on Pd2Ir1, as the energy barriers are generally lower. That suggests that the Pd2Ir2 alloy catalyst is slightly more effective towards the borohydride oxidation reaction. The energy barrier from BOH to B(OH)2 is much lower on Pd2Ir2 than on Pd2Ir1, which means that at a larger concentration of Ir in the alloy, more complete oxidation may be realized. Other than that, the reaction mechanism on both Pd-Ir surfaces followed the same path, for which molecular structures are summarised in Figure 4.13.

Figure 4.14 shows a comparison of the mechanism of reaction of the borohydride oxidation at -1 V vs. Hg/HgO and at -0.64 V vs. Hg/HgO. As seen in the figure, the relative energy of the intermediate species increases and thus are less stable at more negative potentials. The activation energies were also more positive and it can be

128 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

concluded that the reaction mechanism is more favourable at more positive potentials than 1 V vs. Hg/HgO. This agrees with the experimental data, as no current or borohydride oxidation could be measured at -1 V vs. Hg/HgO, which corresponds to the open circuit potential, whereas current densities of around 40 mA cm-2 were measured

-3 -3 at -0.64 V vs. Hg/HgO using a solution containing 0.02 mol dm NaBH4 in 3 mol dm

NaOH.

BH4 BH3 BH2 BH

BOH B(OH)3 B(OH)2 BHOH

Figure 4. 13 Configurations of the preferred borohydride oxidation path on Pd2Ir2(111) surface. The dark blue atoms correspond to Ir, the light blue big atoms correspond to Pd, the pink atoms are B, the red are O atoms and the white are H atoms.

129 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

Borohydride oxidation 1.5

1.01 1 1.70 0.95 0.41 0.49 0.5 0.47 BH4 0 BH3 BH -0.5 0.09 2 BH BHOH B(OH)3 BOH B(OH)2 -1 0.17 BH3 -1.0 V vs. Hg/HgO 0.63 -1.5

/ eV / BH2 -2

-2.5 BH 1.28 -3 -0.64 V vs. Hg/HgO 1.04 -3.5 0.15 -4 BHOH

Relative Energy for borohydride oxidation borohydride for Energy Relative BOH -4.5 B(OH)2 B(OH)3 -5

Figure 4. 14 Preferred reaction mechanism for the borohydride oxidation on Pd2Ir1 at - 1 V vs. Hg/HgO and -0.64 V vs. Hg/HgO.

4.4.3. Competitive borohydride oxidation versus hydrolysis at the Pd-Ir surfaces

During the borohydride oxidation on Pd-Ir surfaces, hydrogen atoms from the

dehydrogenation processes are adsorbed on the electrode surface. Those hydrogen

atoms might either combine with another adsorbed hydrogen atom, forming hydrogen

gas (reaction 4.33), or oxidize with a hydroxyl ion to generate water and release one

electron (reaction 4.34).

⁄ (4.33)

(4.34)

130 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

The free energy of both, hydrogen oxidation and hydrolysis reactions, can be calculated by using equations (4.35) and (4.36), respectively, at 1 atm and 298 K. A borohydride concentration of 0.02 mol dm-3 and 3 mol dm-3 NaOH was considered to calculate the free energy of adsorption.

(4.35)

(4.36)

For an accurate consideration of the competition between hydrogen evolution and hydrogen oxidation, the hydrogen coverage should be considered, as it plays an important role in the hydrogen generation rate. The hydrogen coverage is dictated by the amount of hydrogen released from the borohydride hydrolysis and from the borohydride dehydrogenation occurred during the borohydride oxidation. It can also be related with the borohydride stability and therefore, higher H* coverage will be generated for low alkaline solutions (low pH) or in solutions at higher temperatures.

a) b) c)

Figure 4. 15 Preferred H atom adsorption over: a) Pd(111), b) Pd2Ir1(111) c) Pd2Ir1(111) surfaces at low coverage. The dark blue atoms correspond to Ir, the light blue big atoms correspond to Pd and the white atoms are H atoms.

131 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

For a low hydrogen coverage model, one hydrogen atom was optimized on the Pd-Ir surfaces, giving a hydrogen coverage of 1/9 monolayer (ML) atom for the Pd2-Ir1 and

1/16 ML atom on the Pd2-Ir2 surface. For comparison, the hydrogen adsorption was also studied on a Pd(111) surface. For that purpose, a 4-layer Pd(111) surface was optimized in a 3 × 3 cell. All possible configurations were optimized; demonstrating that the lowest energy for the optimized hydrogen atom was when the adsorbed hydrogen occurred on the fcc configuration at the Pd(111) surface, on the Pd-Ir bridge for

Pd2Ir1(111) and on the top of the Ir atom for Pd2Ir2(111), as shown in Figures 4.15 a), b) and c), respectively.

The comparison of the free energy of the hydrogen generation reaction and hydrogen oxidation reaction versus potential for low hydrogen coverage on Pd(111), Pd2-Ir1 and

Pd2-Ir2 surfaces is shown in Figure 4.16. The free energy of hydrogen evolution did not change with the concentration of Ir in the alloy, for Pd2-Ir1 and Pd2-Ir2 surfaces, with a positive value of 0.43 eV at any potential between -1 V and 1 V vs. Hg/HgO. That suggests that at low hydrogen coverage hydrogen evolution is not favourable (ΔG > 0) in either Pd2-Ir1 or Pd2-Ir2 surfaces, at any studied potential. The free energy of oxidation was positive at potentials from -1 to -0.6 V vs. Hg/HgO, suggesting that the reaction will not spontaneously occur at these potentials. At more positive potentials than -0.6 V vs. Hg/HgO, the H* oxidation starts being favourable (ΔG < 0). At the

Pd(111) surface, both the free energy of hydrogen evolution and oxidation were more positive than on the Pd-Ir surfaces. Therefore, at low H* coverage and at any studied potential, the hydrogen oxidation and evolution are more favourable on the bimetallic surfaces than on pure Pd(111), due to weaker H binding to the surface.

132 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

Figure 4. 16 Reaction free energies for H* oxidation (straight line) and evolution (dash line) of H2 gas as a function of potential for low hydrogen * * coverage 1/9 ML H on Pd2Ir1(111) surface (black lines), 1/16 ML H * on Pd2Ir2(111) surface (red lines) and 1/9 ML H on Pd(111) surface (blue lines). The Pd2Ir1(111) and the Pd2Ir2(111) lines overlay each other.

The binding energy (BE) of the H to the Pd(111), Pd2Ir1(111) and Pd2Ir2(111) surfaces, which can be calculated using equation (4.37) [120], are -0.61 eV, -0.43 eV and -0.41 eV, respectively.

(4.37)

Weaker H binding on the bimetallic surfaces might lead to lower H coverage, and therefore a faster oxidation reaction rate relative to evolution, as the oxidation and evolution rates directly depend on the hydrogen coverage.

133 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

At higher hydrogen coverage, the free energy of evolution and oxidation are expected to decrease to negative values, as increasing the number of hydrogen atoms on the surface makes the easier the reaction to occur. However, at high hydrogen coverage, 9 hydrogen atoms were optimized on the 3 × 3 cell, to give hydrogen coverage of 1 ML atom for the Pd2-Ir1 surface. For the Pd2-Ir2 surface, 16 hydrogen atoms were optimized on the 4

× 4 cell giving 1 ML coverage. The large number of possible configurations of the 9 hydrogen atoms on the surface decreases the likelihood of of locating a global optimum structure. Moreover, the data obtained suffers from model imperfections, as the bimetallic structure disrupts the symmetry and can lead to effects in H binding that are not practical in a typical experimental alloy electrode, which is probably less ordered.

The difference in cell size between Pd2Ir1 and Pd2Ir2 also allows for differences in the surface reorganization allowed. However, these calculations were performed and can provide an orientation of the hydrogen behaviour on the Pd-Ir and Pd surfaces.

Figure 4.17 shows the built and optimized configurations of the nine hydrogen atoms allocated equidistant to each other on the Pd(111), Pd2Ir1(111) and Pd2Ir2(111) surfaces.

At the pure Pd electrode, the optimized structure (d) did not change much from the built one (a), showing that the H atoms are optimized on an fcc configuration. Conversely, on the Pd2Ir1 surface, the 9H changed location during the optimization process from the symmetric hollow sites to be adsorbed on the Ir atoms. At the Pd2Ir2 surface, the 16 hydrogen atoms were initially located on Pd and Ir atop sites, but the hydrogen atoms tend to move towards the Ir instead of the Pd atoms.

134 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

a) d)

b) e)

c) f)

Figure 4.17 H adsorption built for high H coverage over: a) Pd(111), b) Pd2Ir1(111) c) Pd2Ir2(111) surfaces at high coverage. Preferred H adsorption optimized over: d) Pd(111), e) Pd2Ir1(111) f) Pd2Ir2(111) surfaces at low coverage The dark blue atoms correspond to Ir, the light blue big atoms correspond to Pd and the white atoms are H atoms.

The Pd2Ir2 surface seems to offer enough Ir atoms such that all H can move from their symmetric hollow sites towards Ir atoms, making stripes of H atoms, as shown in Figure

135 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

4.17f). These Ir stripes, which form a regular pattern, might provide a greater stabilization to the H high coverage and the strong H binding could hinder the selectivity of the catalyst towards the oxidation. However, it should be mentioned that these calculations consider an ideal material configuration consisting of Pd layers followed by Ir layers that are unachievable experimentally.

The results show more negative and thus more favourable free energies of hydrogen oxidation and hydrogen evolution, than that obtained at low hydrogen coverage, as it was expected. At high hydrogen coverage on the Pd2Ir1 surface, the free energy of the hydrogen evolution was favourable (approximately -0.73 eV) at all potentials between -

1 and 1 V vs. Hg/HgO. That value was 0.7 eV more negative on Pd2Ir2, suggesting that the presence of Ir makes more favourable the hydrogen generation. The results also showed that on Pd2Ir1 and Pd2Ir2 surfaces, the free energy of the hydrogen oxidation constantly decreased towards more negative values at all potentials. The crossing point at which oxidation competes with evolution does not vary for the studied materials (Pd,

Pd2Ir1 and Pd2Ir2). The two lines corresponding to free energy of hydrogen evolution and oxidation, intercept at approximately -1 V vs. Hg/HgO, being the free energy of oxidation lower than the free energy of hydrogen evolution at more positive potentials.

This suggests that at more positive potentials than -1 V vs. Hg/HgO, the hydrogen atom adsorbed on the surface would oxidize rather than evolve as hydrogen bubbles. At more negatives potentials than -1 V vs. Hg/HgO, the hydrogen evolution is more favourable than the oxidation and thus, hydrogen bubbles would evolve. This agrees with the experimental results showed in Figure 4.1, as hydrogen generation was observed at the open circuit potential (-0.99 V vs. Hg/HgO) and at more positive potentials, the hydrogen generation substantially decreased. It can be concluded that the presence of Ir

136 Chapter 4: Pd-Ir coated microfibrous carbon for borohydride oxidation

in the bimetallic alloy decreases the free energy of oxidation, making the Pd2Ir2 surface more favourable for the borohydride oxidation at all potentials, compared to the Pd2Ir1.

4.5. Conclusions

The catalytic activity of the Pd-Ir alloy towards the borohydride oxidation and its hydrolysis was experimentally and computationally analysed. The Pd-Ir coated microfibrous carbon supported on a titanium plate electrode showed catalytic activity towards the borohydride oxidation, obtaining current density values from 100 to 200

-2 -3 -3 mA cm , when a solution containing 0.5 mol dm NaBH4 in 3 mol dm NaOH and a 3 cm2 electrode geometric surface area, were used. At the same time, the hydrogen generation rate was minimised to less than 0.1 dm3 min-1. DFT analysis showed that the presence of Ir seems to favour the borohydride oxidation, as lower activation barriers were obtained on the Pd2Ir2 alloy than on the Pd2Ir1 alloy. The results from the DFT calculations also showed that in the presence of Ir, the Gibbs free energy of oxidation is lower than that on pure Pd, at all potentials, which suggests that the oxidation reaction starts at more negative potentials on the alloy than on pure Pd.

The use of DFT calculations can benefit the research on DBFC, as they can be very helpful in the design and previous characterization of catalysts, including pure metals or bimetallic alloys. The mechanism of reaction of borohydride can be predicted and the competition between the hydrogen oxidation and evolution can be studied.

Experimental work using different catalysts can be avoided and only carried out on catalytic materials that previously showed good performance according to the DFT analysis.

137 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

In order to minimize the hydrogen generation rate during the oxidation of borohydride ions in alkaline solution and improve the performance of the direct borohydride fuel cell, several surfactants, including Triton X-100, Zonyl FSO, S-228M, sodium dodecyl sulphate (SDS) and FC4430 were evaluated. Electrolysis at constant current or constant

-3 -3 o potential in solutions containing 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C in the absence and the presence of these additives was performed using two different working electrodes: 1) a planar gold and 2) a dispersed nanoparticulate gold supported on carbon

(Au/C). The rate of hydrogen generation, current density and anode potential were measured during the electrolysis.

After the effect of surfactants on the borohydride hydrolysis was evaluated, the kinetics of the borohydride oxidation in the presence of surfactants was analysed using cyclic voltammetry (CV) and linear sweep voltammetry at a rotating disc electrode (RDE).

-3 -3 Cyclic voltammetry of solutions containing 0.02 mol dm NaBH4 in 3 mol dm NaOH, at 23 oC, in the absence and in the presence of surfactants including Triton X-100, SDS and FC4430 was carried out on the two working electrodes: a planar gold rotating disc and Au/C. The diffusion coefficient of borohydride ion, heterogeneous rate constant for borohydride oxidation, standard rate constant and charge transfer coefficient were calculated in the absence and in the presence of surfactants.

5.1. Electrolysis at constant current

Table 5.1 summarises the rate of gas generated during electrolysis at 200 mA cm-2 using a 3 cm2 planar Au and dispersed Au/C electrodes. All solutions contained 1 mol dm-3

138 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

-3 NaBH4 in 3 mol dm NaOH and solutions b) to d) included: b) 0.00001 wt. %, c)

0.0001 wt. %, and d) 0.001 wt. % of the surfactants SDS, Triton X-100, Zonyl FSO and

S-228M.

Elect. Hydrogen generation rate / cm3 min-1 Sol. Planar Au electrode Dispersed Au/C electrode Triton Zonyl Triton Zonyl S- SDS S-228M SDS X-100 FSO X-100 FSO 228M a) 2.4 ± 0.2 1.3 ± 0.2

2.9 ± 2.4 ± 3.1 ± 2.9 ± 1.2 ± 1.2 ± 1.3 ± 1.4 ± b) 0.2 0.2 0.1 0.2 0.1 0.1 0.1 0.1 3.4 ± 2.9 ± 3.0 ± 2.9 ± 1.3 ± 1.1 ± 1.3 ± 1.5 ± c) 0.1 0.1 0.3 0.2 0.3 0.2 0.1 0.1 3.2 ± 3.2 ± 2.9 ± 2.8 ± 1.2 ± 1.0 ± 1.4 ± 1.4 ± d) 0.1 0.1 0.2 0.5 0.3 0.2 0.1 0.1

Table 5.1 Hydrogen generation rate during constant current (600 mA) electrolysis - 2 of BH4 at a planar Au and an Au/C electrode (each 3 cm ). The solutions -3 -3 contained 1 mol dm NaBH4 in 3 mol dm NaOH and the following concentration of surfactants: a) 0 wt. %, b) 0.00001 wt. %, c) 0.0001 wt. %, d) 0.001 wt. %. The surfactants used were: SDS, Triton X-100, Zonyl FSO and S-228M.

The electrode potential during the electrolysis of these solutions varied between -0.5 and 0 V vs. Hg/HgO, this region corresponding to borohydride oxidation. However, the electrolysis also produced gas from the secondary reaction, the hydrolysis of

- borohydride. The composition of gas generated in the solution containing BH4 and no surfactants (solution a) in Table 5.1) was analysed by gas chromatography giving 92.3

% vol. hydrogen and 7.4 % vol. nitrogen. Therefore, it was assumed that the only gas generated during the electrolysis in the presence of borohydride was hydrogen from the incomplete oxidation of borohydride and its hydrolysis and that the electrolysis of water

139 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

did not take place as no oxygen was found. This was confirmed by the electrolysis of a

3 mol dm-3 NaOH solution, where the electrode potentials increased to a range between

1.50 and 1.74 V vs. Hg/HgO that corresponds to oxygen evolution in an alkaline system; this was more positive than the potential range observed during electrolysis in the presence of borohydride. The small percentage of nitrogen measured in the sample was due to the use of this gas for cleaning and purging the sampling gas bags before the experiment.

The volumetric rates of gas generated at the planar gold and the dispersed nanoparticulate gold supported on carbon (Au/C) electrodes during the electrolysis of the borohydride solution in the absence of surfactant were 2.4 ± 0.1 cm3 min-1 and 1.3 ±

0.1 cm3 min-1, respectively. The values are of the same order of magnitude as those obtained by Wang et al. [95], who reported 1.89 cm3 min-1 at 600 mA using a 4 cm2

-3 -3 Au/C 20 wt. % electrode from a solution containing 0.7 mol dm NaBH4 in 2 mol dm

NaOH. The effect of surfactants on the hydrogen generation rate is discussed below for both working electrodes.

5.1.1. Planar Au electrode

The rates of gas generated in the absence and in the presence of surfactants are compared in Table 5.1. The rate of hydrogen gas generated at the Au planar electrode slightly increased with the addition of 0.0001 wt. % SDS to the electrolyte solution, reaching a maximum of 3.4 ± 0.1 cm3 min-1. In the presence of 0.00001 wt. % Triton X-

100, the hydrogen generation rate remained the same as that without surfactants. When the concentration of Triton X-100 in the electrolyte solution was increased to 0.0001 wt.

140 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

%, the hydrogen generation rate increased by 15 % compared to that without surfactants, and to 24 % when the concentration was further increased to 0.001 wt. %.

The presence of 0.00001 wt. % and 0.0001 wt. % S-228M also increased the hydrogen generation rate by 18 % respect to that without surfactant. An increase of around 0.7 cm3 min-1 was obtained in the presence of Zonyl FSO, compared to that in the absence of surfactants. This increase did not seem to be influenced by the concentration over the studied range.

The presence of the surfactants tested at concentrations from 0.00001 wt. % to 0.001 wt.

% seems to favour the hydrolysis of borohydride as it increased the hydrogen generation rate at the planar gold electrode. However, their effect on the hydrogen generation rate is very small and higher concentrations of surfactants were analysed to observe a more obvious effect on the hydrogen generation rate.

5.1.2. Dispersed Au/C electrode

As shown in Table 5.1, when 0.00001 wt. % and 0.0001 wt. % Triton X-100 were added to the electrolyte, the hydrogen generation rate decreased by 6 % and 12 %, respectively, compared to that in the absence of surfactant. Increasing the concentration of Triton X-100 to 0.001 wt. %, a further decrease of 23 % was observed. The presence of 0.001 wt. % S-228M, realized an increase of 11 % with respect to the hydrogen generation rate without surfactant and a more noticeable increase of 27 % was observed in the presence of 0.001 wt. % SDS. In contrast, the hydrogen generated did not seem to be highly influenced by the presence of Zonyl FSO, with a maximum increase of 7 % at the maximum concentration of 0.001 wt. %.

141 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Since 0.001 wt. % Triton X-100 decreased the H2 generation by 23 %, this surfactant might help to increase the fuel utilization of the DBFC. However, further experiments need to be carried out to measure whether or not the current density from the borohydride oxidation is inhibited along with the hydrolysis reaction when this surfactant was used.

5.2. Potentiostatic electrolysis

5.2.1. Planar Au electrode

As the concentration of surfactants used in the previous experiments seemed to have very little effect on the rate of hydrogen gas generated from the hydrolysis of borohydride, the concentration of surfactants was increased to 0.1 wt. %. The surfactant

Zonyl FSO, which moderately increased the hydrogen generation rate at the planar Au electrode and barely varied the hydrogen generated at the Au/C electrode, suggested

- that this surfactant has no benefit for the BH4 oxidation and was not further evaluated.

Instead, the non-ionic fluorinated surfactant FC4430, which showed promising results in preliminary experiments carried out in our laboratory, was studied.

Figure 5.1 shows the variation of the hydrogen generation rate at different applied

-3 -3 electrode potentials using a solution of 1 mol dm NaBH4 in 3 mol dm NaOH, in the absence and in the presence of surfactants: a) no surfactant, b) 0.1 wt. % of Triton X-

100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430. In all cases, the rate of gas generation increased when a more positive potential was applied, but the presence of each surfactant had a different effect on the borohydride reactions.

142 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Figure 5.1 Gas generation rate vs. potential for a planar Au electrode. All the -3 -3 o solutions contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C and the presence of the surfactant was varied: a) without surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430.

Triton X-100 (curve b) and SDS (curve c) increased the hydrogen generation rate by an average of 50 % and 25 %, respectively, compared to that in the absence of surfactants, at all potentials. In contrast, the addition of FC4430 decreased the hydrogen generation rate, at all applied potentials, compared to that in its absence. At 0.2 V vs. Hg/HgO and concentrations of 0.1 wt. % (curve d) and 0.3 wt. % FC4430 (curve e), the hydrogen generation rate decreased by 37 % and by 50 %, respectively.

Figure 5.2 illustrates the current density measured during the electrolysis as a function of the applied potential. In general, the current density slightly varied with the addition of surfactants during the electrolysis at negative potentials but it noticeably increased

143 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

during the electrolysis at positive potentials. Only the presence of Triton X-100 seems to increase the current density at all applied potentials, and more noticeably from potentials more positive than -0.2 V vs. Hg/HgO. This increase in current density is probably due to the hydrogen oxidation, since more hydrogen was generated in the presence of Triton X-100, as shown in Figure 5.1. When 0.1 wt. % FC4430 was present in the solution, the current density remained close to that without surfactants, which suggests that the FC4430 inhibits the hydrolysis but it does not affect the borohydride oxidation. At positive potentials vs. Hg/HgO the current density becomes higher than that with no surfactant. At higher concentrations of FC4430, 0.3 wt. %, the current density slightly decreased in comparison with that in the presence of 0.1 wt. % FC4430, being the decrease more pronounced at potentials between -0.15 V vs. Hg/HgO to 0.05

V vs. Hg/HgO, as shown in Figure 5.2.

At -0.15 V vs. Hg/HgO, the current density was 10 % lower in the presence of 0.3 wt. %

FC4430 than in the presence of 0.1 wt. % FC4430. It is probable that at 0.3 wt. %

FC4430, a layer of adsorbed surfactant molecules has been formed on the electrode

- surface. At more positive potentials than 0.05 V vs. Hg/HgO, BH4 ions may be able to penetrate this layer to be oxidised at the electrode surface, increasing the current density.

144 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Figure 5.2 Current density vs. electrode potential at the planar Au anode. All the -3 -3 o solutions contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C. Some of them also contained surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430.

Figure 5.3 shows the hydrogen generation rate versus current density obtained during

-3 the electrolysis of borohydride ions in solutions containing 1 mol dm NaBH4 in 3 mol dm-3 NaOH in the presence and in the absence of surfactants, at the planar Au electrode.

The figure also shows straight lines, which indicate the apparent number of electrons involved in the borohydride oxidation according to equation (4.1) proposed by Wang et al. [95] in order to calculate the number of electrons released during the electrolysis. As seen in Figure 5.3 (line a), the number of electrons released during the electrolysis of the solution in the absence of surfactants is 4, and this number increased in the presence of surfactants, being between 5 and 6 in the presence of 0.1 wt. % FC4430 (line d). The increase in the number of electrons released in the presence of SDS and Triton X-100 is

145 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

probably due to the oxidation of hydrogen, as more hydrogen is released in the presence of these surfactants. The reason why the number of electrons obtained is 4 instead of 8 might be because the range of potentials used for the electrolysis corresponds to the kinetic control region instead of the mass transfer plateau.

14 -

2e - 1

- 4e min 12 3 a) b) 10 c)

8 d) 6

4 6e-

2 Hydrogen gas generation rate / cm /rate gas generation Hydrogen 8e- 0 0 100 200 300 400 500 600 700 Current density, j / mA cm2

Figure 5.3 Hydrogen generation rate vs. current density at the planar Au anode. All -3 -3 o the solutions contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C. Some of them also contained surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS and d) 0.1 wt. % FC4430.

In summary, the presence of Triton X-100 and SDS increased the current density and the hydrogen generation rate. This trend was already observed at lower concentrations

(0.001, 0.0001 and 0.00001 wt. %) of Triton X-100 and SDS in the electrolysis of borohydride ions at 200 mA cm-2 constant current electrolysis as shown in Table 5.1.

For that reason, both surfactants might be useful for an IBFC using a gold flat plate anode, where higher hydrogen generation rates are required. In contrast, the use of

146 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

surfactant FC4430 might increase the fuel utilization and the performance of the DBFC, when using the planar gold electrode. This surfactant decreased the hydrogen generation rate by 50 % compared to that in the absence of surfactant; however, it also decreased the current density by an average of 30 %. By increasing the concentration of FC4430 the gas generation decreased and the current density remained approximately the same as that in the absence of surfactants. This suggests that the addition of an optimum concentration of FC4430 to the alkaline solution could minimise the hydrogen generation without decreasing in the current density substantially.

5.2.2. Au/C Electrode

Figure 5.4 shows the hydrogen generation rate vs. current density obtained during the

-3 electrolysis of borohydride ions in a solution containing 1 mol dm NaBH4 in 3 mol dm-3 NaOH, at the Au/C electrode (10 wt. % and loading 0.5 mg cm-2). The figure also shows the straight lines corresponding to the number of electrons 2, 4, 6 and 8, calculated using equation (4.1) proposed by Wang et al. [95] and previously shown in

Figure 5.3. Figure 5.4 allows the analysis of the experimental results and their comparison with those previously reported by Wang et al. [95] who elucidated the mechanism of reaction by plotting the H2 generation rate versus the current.

According to Figure 5.4, the number of electrons transferred during the electrolysis of 1

-3 -3 -2 mol dm NaBH4 in 3 mol dm NaOH was 8 for current densities up to 40 mA cm and approximately 6 for higher current density values. Figure 5.5.a) shows a plot of the hydrogen generation rate vs. electrode potential during the electrolysis of 1 mol dm-3

-3 NaBH4 in 3 mol dm NaOH and Figure 5.5.b) shows the apparent number of electrons

147 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

released during the electrolysis, calculated from equation (4.1) compared with the applied potential.

8 -

2e -

1 4e

- 7

min

3 6

4 6e- 3

Hydrogen gas generation rate / cm /rate gas generation Hydrogen 1 8e- 0 0 200 400 600 Current density, j / mA cm2

Figure 5.4 Hydrogen gas generation rate vs. current density in a solution containing -3 -3 o 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C using an Au/C 10 wt. % electrode.

In Figure 5.5.a) the slope of the hydrogen generation rate changes with the electrode potential; when the slope was zero (E < -0.4 V vs. Hg/HgO), the oxidation of borohydride was complete with 8e- released and no hydrogen generation. At more positive potentials, between -0.4 V and 0 V vs. Hg/HgO, the hydrogen generation rate increases with the potential and reaches a maximum at 0 V vs. Hg/HgO, after that there is a sudden decrease between 0 and 0.1 V vs. Hg/HgO. At more positive potentials, the slope of the gas generation rate versus potential increased again and it seems to recover following the trend observed until 0 V vs. Hg/HgO.

148 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

1 -

min

3 6

1 Hydrogen gas generation rate / cm /rate gas generation Hydrogen 0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 Electrode Potential, E vs. Hg/HgO / V

a) 8

app n 6

Apparent number of electrons,of number Apparent 1

0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 Electrode potential, E vs. Hg/HgO / V

b) Figure 5.5 a) Hydrogen gas generation rate vs. electrode potential and b) apparent number of electrons released during oxidation vs. electrode potential -3 from the electrolysis of a solution containing 1 mol dm NaBH4 in 3 mol dm-3 NaOH at 23 oC using an Au/C 10 wt. % electrode.

There are two possible explanations for the minimum appreciated at 0.1 V vs. Hg/HgO.

The first possibility is that a higher free energy of adsorption of borohydride at that

149 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

potential leads to less borohydride adsorption, lower hydrogen generation rate and lower current. At more positive electrode potentials than 0.2 V vs. Hg/HgO, a lower free energy of adsorption allows a higher borohydride adsorption, which provides higher current again. The second possibility is the formation of gold or boron oxides at that potential, which passives the electrode and inhibits the borohydride oxidation. The experimental details support this possibility, as the current density measured at 0.1 V vs.

Hg/HgO decreased with time as the electrolysis was being performed, suggesting that there is a deposition on the electrode that is passivizing the electrode and reducing the borohydride oxidation. At 0.2 V vs. Hg/HgO the electrode has the same activity as it did at potentials more negative than 0 V vs. Hg/HgO.

As it can be appreciated in Figure 5.5.b), the number of electrons between -0.5 and -0.4

V vs. Hg/HgO approaches 8 and, as the potential becomes more positive, the number of electrons decreases to a value between 5 and 6.

Figure 5.6 shows the CV of the borohydride oxidation on the Au/C electrode. The first peak a2, observed at -0.4 V vs. Hg/HgO has been reported to be due to the direct oxidation of borohydride, where the number of electrons is meant to be 8. A second oxidation wave a3, is a continuation of peak a2 and has been reported to be due to the oxidation of intermediate species, and thus the number of electrons released must be less than 8 and a higher hydrogen generation rate is observed. The results shown in

Figure 5.5.b) agree with the interpretation of the typical CV of the borohydride

- oxidation, where at more negative potentials the number of electrons is 8 (direct BH4 oxidation) and at more positive potentials, the hydrogen generation due to the hydrolysis lead to a number of electrons below 8, according to Figure 5.5.b) 6 e-. Figure

5.6 however, does not show the typical current decrease that is normally appreciated at

150 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

about 0.4 V vs. Hg/HgO in the cyclic voltammetry of borohydride oxidation on planar

Au surfaces. That decay of the current density is attributed to the electrode surface being covered by either hydrogen bubbles from the borohydride hydrolysis or borohydride oxidation by-products. The hydrogen generation on the Au/C surface is much lower than that on the planar gold surface and the hydrogen bubbles generated might not be enough to cover the electrode surface and decrease the current.

2 - 25

/ mA cm mA /

, j ,

15 a3

10 Current density Current 5

0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 Electrode Potential, E vs. Hg/HgO / V

- 2 Figure 5.6 Cyclic voltammogram for BH4 oxidation at a Au/C RDE (0.5 cm ) in the absence of surfactants at 23 oC.

Chatenet et al. [56] suggested that the borohydride oxidation follows different pathways at low and at high electrode potentials. At low potentials (E < 0.3 – 0.5 V vs. RHE) the

− generation of BH3OH and H2 occurs according to reaction (2.8) followed by the

− oxidation of BH3OH (reaction (2.13) or (2.14)), releasing from three to six electrons.

They obtained hydrogen generation at the open circuit potential, which oxidises at negative potentials and sharply increases the current due to the oxidation of

151 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

borohydride, peak a2. At more positive potentials (E > 0.3 – 0.5 V vs. RHE), the reaction pathway involves the oxidation of borohydride ions through reaction (2.12) followed by reaction (2.13) or (2.14), with a total release of between five and eight electrons. In this work no hydrogen was generated at the open circuit potential, due to the use of a more alkaline solution (3 mol dm-3 NaOH versus 1 mol dm-3 NaOH used by

Chatenet et al.). The trend of the curve in Figure 5.5.a) also agrees with that reported by

Wang at al. [95], who performed potentiostatic electrolysis on an Au/C electrode (20 wt. % Au) and used solutions containing 0.3 mol dm-3, 0.5 mol dm-3 and 0.7 mol dm-3

-3 NaBH4 in 2 mol dm NaOH. A maximum hydrogen generation rate was also obtained; however, it was shifted to more positive values by approximately 0.67 V compared to that reported in this work. This difference in potential might be attributed to the concentration of the NaOH solution (3 mol dm-3 NaOH in this work compared to 2 mol dm-3 used by Wang et al.) and the different Au loadings (20 wt. % vs. 10 wt. % used in this work). According to the results published by Wang et al., a hydrogen generation rate of 0.5 cm3 min-1 was already being observed at the open circuit conditions, -1.15 V vs. Hg/HgO. However, no gas generation was observed at zero current (open circuit voltage) in any experiment carried out in the present work, at either the planar Au or the

Au/C electrode. This is probably due to the use of a higher concentration of NaOH in this work. Wang et al. attributed the changes obtained in the polarization curves to surface changes, but provided no further explanation. Figure 5.7 shows the gas generation at the Au/C electrode during the electrolysis of solutions with and without surfactants at different potentials.

152 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Figure 5.7 Rate of gas generation vs. potential applied to an Au/C electrode. The -3 -3 o solutions contained 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C. a) without surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430.

Generally, the presence of any surfactant led to less hydrogen evolution. Below 0 V vs.

Hg/HgO, FC4430 decreased the gas generation by an average of 15 % and 45 % for concentrations of surfactant of 0.1 wt. % and 0.3 wt. %, respectively. The presence of

0.1 wt. % Triton X-100 caused a decrease of 50 % in the rate of hydrogen generation at

0.4 V vs. Hg/HgO, from 6.8 cm3 min-1 to 3.3 cm3 min-1, as shown in Figure 5.7.

In the presence of 0.1 wt. % SDS, the gas generation rate decreased by an average of 4

% when potentials from -0.4 V vs. Hg/HgO to -0.1 V vs. Hg/HgO were applied. At 0 V vs. Hg/HgO, the gas generation decreased about 30 %, but increased almost the same percentage at 0.1 V vs. Hg/HgO. The transition region observed in the absence of

153 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

surfactant, between 0 and 0.2 V vs. Hg/HgO was slightly shifted to less positive potentials, between -0.1 and 0.1 V vs. Hg/HgO, when 0.1 wt. % SDS was in the solution. It is believed that the shift in the transition region compared to that obtained in the absence of surfactants is related with the free energy of adsorption of SDS, which reaches a minimum at around 0 V vs. Hg/HgO. At that potential, the SDS molecules are highly adsorbed, forming a layer of surfactant on the electrode [188], which hinders the borohydride oxidation and its hydrolysis and might also affect the potential at which gold oxide is formed.

It is noticeable that when Triton X-100 or FC4430 were in the solution, the slope of the hydrogen generation rate vs. potential did not change as it did in the absence of surfactants or in the presence of SDS. In the presence of Triton X-100 and FC4430 surfactants, the minimum in Figure 5.7 curve a) due to the formation of gold oxide, the layer of adsorbed surfactants probably passives the gold surface protecting it from the oxidation.

Figure 5.8 illustrates the current density obtained during the electrolysis at the Au/C electrode and different constant potentials of solutions in the absence and in the presence of surfactants. The oxidation current decreased nearly 60 % in the presence of

0.1 wt. % FC4430, and between 70 - 80 % in the presence of 0.3 wt. %. The oxidation/ hydrolysis reaction started at -0.4 V vs. Hg/HgO, which is 0.1 V vs. Hg/HgO more positive compared with the oxidation onset of the solution without surfactants, suggesting that in the presence of the FC4430 the reactions require more energy to start.

The presence of SDS not only decreased the hydrogen generation rate slightly but also the current density.

154 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Figure 5.8 Polarization curves from experiments applying a constant potential at the 2 -3 Au/C cloth electrode (3 cm ). Solutions contained 1 mol dm NaBH4 in 3 mol dm-3 NaOH at 23 oC: a) without surfactant, b) 0.1 wt. % Triton X- 100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430.

5.2.3. Triton X-100

At the Au/C electrode, the hydrogen generation rate generally decreased in the presence of 0.1 wt. % Triton X-100 compared to that without surfactants, i.e. at 0.4 V vs.

Hg/HgO, from 6.8 cm3 min-1 to 3.3 cm3 min-1, as previously shown in Figure 5.7. The presence of 0.1 wt. % Triton X-100 also decreased the current density compared with that without surfactants, as shown in Figure 5.8, i.e. 65 % decrease at 0.2 V vs.

Hg/HgO. That would indicate that the addition of Triton X-100 at 0.1 wt. % is not beneficial for the borohydride oxidation. However, the addition of lower concentrations

(0.00001 wt. %, 0.0001 wt. % and 0.001 wt. %), decreased the gas generation rate during electrolysis by 6 %, 12 % and 23 % respectively, compared to that in the absence of surfactant, as previously reported in Table 5.1. These results suggest that there is an

155 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

optimum concentration of Triton X-100, at which the hydrolysis is inhibited without affecting the oxidation rate. At higher concentrations that the optimum, the surfactant limits the number of active sites on the electrodes avoiding the access and inhibiting the borohydride hydrolysis together with its oxidation. This optimum concentration may be related to the critical micelle concentration (CMC), which is characteristic for each surfactant and commonly varies from 10-4 mol dm-3 to 10-2 mol dm-3. The CMC of

Triton X-100 has been reported between 0.01 wt. % and 0.03 wt. % (1.65 × 10-4 - 5 ×

10-4 mol dm-3) in water [228].

Figure 5.9 shows current density vs. potential in the absence of surfactants and in the presence of concentrations of Triton X-100 above and below the CMC (0.1 wt. % and

0.001 wt. %, respectively). At concentrations below the CMC, such as 0.001 wt. %, the current density remains the same as that in the absence of surfactants at potentials lower than -0.25 V vs. Hg/HgO. At more positive potentials, the current density increased by the addition of 0.001 wt. % Triton X-100, because of a more complete borohydride oxidation. At that concentration, the surfactant is adsorbed on the electrode, in monolayers, and inhibits the borohydride hydrolysis without the inhibition of the borohydride oxidation, perhaps allowing the diffusion of borohydride through the layer when the potential increases slightly. The current density in the presence of 0.1 wt. %

Triton X-100 is lower than that in the absence of surfactants, probably because at concentrations higher than the CMC (e.g. 0.1 wt. % Triton X-100), the micelles are adsorbed in the electrode hindering the access of borohydride ions to the electrode surface and thus, decreasing the borohydride oxidation and the current density. The presence of Triton X-100 decreased the current density compared to that in the absence of surfactant, i.e. 65 % decrease at 0.2 V vs. Hg/HgO. This suggests that only at the

156 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

optimum concentration, Triton X-100 may be a good hydrolysis inhibitor. The following section gives a more detailed explanation of the behaviour of Triton X-100 towards the borohydride hydrolysis.

600

500 0.001 wt. %

-2 Triton X-100 400 No surfactant

, mA cm , mA

j 300

200

Current density density Current

0.1 wt. % 100 Triton X-100

0 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 Electrode potential, E vs. Hg/HgO / V

Figure 5.9 Hydrogen generation rate vs. current density during electrolysis of -3 -3 o solutions containing 1 mol dm NaBH4 in 3 mol dm NaOH at 23 C: a) without surfactants, b) with 0.001 wt. % Triton X-100, c) with 0.1 wt. % Triton X-100.

Figure 5.10 illustrates the adsorption conformations of surfactant molecules on solid surfaces depending on their concentration in the solution. An electrode surface in contact with a diluted surfactant solution will have individually adsorbed surfactant molecules (Figure 5.10 a). When the concentration of surfactant increases to concentrations of approximately 0.1 times the CMC (Critical Hemi-Micellar

Concentration, CHMC), the surfactants start to aggregate forming the called

“hemimicelles” in Figure 5.10 b), which might be adsorbed on the electrode surface

[189].

157 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Electrode Electrode surface surface

Monomers HemiMicelles C < CMC CHMC > C > CMC a) b)

Electrode Electrode surface surface

Monolayer Bilayer C > CMC C >> CMC c) d)

Figure 5.10 Different states of the surfactant molecules depending on the concentration: a) C < CMC Surfactant molecules adsorbed as monomers on the electrode, b) CHMC > C > CMC surfactant molecules adsorbed in form of hemimicelles, c) C > CMC monolayers of surfactants adsorbed on the electrode, d) C >> CMC, bilayers of surfactant monomers adsorbed on the electrode.

Above CHMC, the adsorption of surfactant increases rapidly, forming the surface bilayer almost instantaneously when the CMC is reached, as shown in Figure 5.10 c).

When the concentration of surfactant is further increased, the number of micelles increases until a second transition, where the aggregation number of the micelles and

158 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

their shape change [189]. Increasing the concentration, the surface is rapidly covered by a surfactant bilayer, and no further adsorption takes place (Figure 10 d). The absorption saturation plateau occurs when the solution concentration is at or above the CMC [189].

The adsorption and the acceleration of the dissolution process of the surfactant are affected by the change of temperature.

In the case of Triton X-100, at concentrations below the CMC, such as 0.001 wt. %, the surfactant is adsorbed on the electrode, in monolayers inhibiting the borohydride hydrolysis but not the borohydride oxidation, perhaps allowing the diffusion of borohydride through the layer when the potential increases slightly. However, at concentrations higher than the critical micelle concentration (e.g. 0.1 wt. %), the micelles might be adsorbed in the electrode hindering the access of the borohydride ions to the electrode surface and thus, decreasing the current density due to the borohydride oxidation.

In order to analyse how the Triton X-100 molecule is adsorbed on the surface electrode and how that adsorption affects the borohydride oxidation, the results of DFT studies are considered in the following section.

5.2.3.1. DFT study of the adsorption of Triton X-100 on Au(111) and its effect on borohydride oxidation

DFT calculations were used to study the adsorption of a Triton X-100 molecule on an

Au(111) electrode surface and its effect in the borohydride adsorption energies. All calculations were performed using the ab initio total energy and the VASP program

159 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

[173-175] following the same procedure previously described in Chapter 4. The

Au(111) crystalline structure was modelled using a 4 layer slab on a 3 × 6 surface cell, with the two top layers relaxed and the two bottom layers constrained to face-centred- cubic (fcc) lattice position. The cell was modelled from a 4 layer slab on a 3 × 3 cell previously optimized by Rostamikia et al. [119]. The size of the initial cell (3 × 3) had to be doubled in one of the sides due to the big size of the Triton X-100 molecule. A

Monkhorst-Pack grid [182] was used to optimize the structure followed by a 5 × 5 × 1 grid single-point calculation to give the total energy. A vacuum of 13 Å was inserted between the periodic slab to represent the electrode surface and analyse the species interactions with the electrode surface. The projected augmented wave method and a plane wave basis set (cut-off energy 450 eV) were employed to calculate the wave function [183, 184].

Several possible configurations of the Triton X-100 molecule and of the BH4 molecule were optimized separately on the Au(111) surface. In order to analyse the effect of the surfactant molecule in the borohydride oxidation, the adsorption energies of Triton X-

- - 100, BH4 and BH4 in the presence of an adsorbed Triton X-100 molecule, were calculated assuming the lowest energy configuration. The Triton X-100 molecule, which structure is:

CH3 CH3 (5.1) H3C-C-CH2-C O-(CH2CH2O)N-H

CH3 CH3

With N ≈ 9.5. In order to simplify the calculations a value of N = 1 was assumed. With this consideration, the molecule length is reduced allowing the calculation to be finished in a reasonable period of time. The length of the molecule should not affect the

160 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

preferred adsorption structure, and it can provide an approximation of the molecule adsorption configuration.

The adsorption energy versus the electrode potential was calculated and compared for all the different molecule configurations studied in order to find out the optimum Triton

X-100 molecule arrangement for the adsorption on the gold surface. The adsorption energy of Triton X-100 over the Au(111) crystalline surface, reaction (5.2), was calculated using equation (5.3):

(5.2)

(5.3)

Where * represents the bare surface, is the energy of adsorption of Triton X-100

on the surface of the electrode, is the free energy of the molecule in aqueous

solution, is the free energy of the adsorbed species, e is the charge on an electron, U is the absolute (vacuum reference) electrode potential and is the free energy of the bare surface.

The two preferred Triton X-100 adsorption configurations constrained an OH bond breaking, where the hydrogen atom is adsorbed separately and the rest of the molecule is adsorbed from the oxygen atom. This configuration is reasonable considering the high tendency of the oxygen atoms to be adsorbed on gold surfaces. Figure 5.11 shows the preferred (lowest in energy) configuration of Triton X-100, vertically adsorbed on

Au(111), with a borohydride molecule in the vicinity.

161 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Triton X-100*

* BH4

a) b)

Triton X-100*

* BH4

c) d)

Figure 5.11 Atomic structure of the Triton X-100 molecule next to a borohydride molecule on Au(111): a) vertically adsorbed side view, b) vertically adsorbed front view, c) horizontally adsorbed side view, d) horizontally adsorbed front view. The yellow atoms correspond to Au, the grey atoms correspond to C, the pink atoms are B, the red are O atoms and the white are H atoms.

The CH3 groups at the top of the Triton X-100 molecule were substituted by H groups for simplification and for the molecule to fit in the 13 Å vacuum. The configuration of the adsorbed borohydride molecule did not change by the presence of the Triton X-100 molecule, the hydrogen atoms were adsorbed on the gold atoms atop site and the boron atom was on the hollow site, which was the same configuration than when the

162 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

borohydride molecule was adsorbed in the absence of Triton X-100.The free energy of adsorption of borohydride on the Au(111) surface in the absence and in the presence of the Triton X-100 was calculated using equation (5.4):

(5.4)

where and are the free energy of the borohydride ions in aqueous solution and adsorbed on the surface, respectively. The calculation of the free energy of adsorption of borohydride on the gold surface in the presence of Triton X-100, the free energy of the bare surface was considered as the free energy of the Triton X-100 adsorbed. Figure 5.12 shows a comparison of the free energy of adsorption of borohydride ions in the absence and in the presence of the adsorbed Triton X-100 molecule versus potential. As it can be appreciated in the figure, the adsorption of borohydride without Triton X-100 starts being favourable at around -0.66 V vs.

- Hg/HgO (-0.8 V vs. NHE), whereas in the presence of Triton X-100 the BH4 adsorption is favourable at any potential from -1 to 1 V vs. Hg/HgO.

163 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

0.5

0 No surfactant

Horizontally adsorbed -0.5 Triton X-100

-1

Vertically adsorbed -1.5 Triton X-100

-2 Free energy of adsropetion , ∆Gads /eV ∆Gads , adsropetionof energy Free -2.5 -1 -0.5 0 0.5 1

Electrode potential, E vs. Hg/HgO / V

Figure 5.12 Comparison of the free energy of adsorption of borohydride ions on Au(111) in the absence and the presence of the two preferred configurations for the adsorbed Triton X-100 molecule vertically and horizontally.

The free energy of adsorption of borohydride decreased by 0.42 eV, when Triton X-100 was adsorbed on the electrode. This suggests that the presence of Triton X-100 favours

- the BH4 adsorption and thus the oxidation reaction. In this study low coverage of surfactant was considered however, at high coverage (high concentration) it is probable that the surfactant will have a different effect on the adsorption energy of the borohydride ions. These conclusions were also observed experimentally; when it was found that, at low concentrations, Triton X-100 increased the current density and decreased the hydrogen generation from the hydrolysis.

In conclusion, although these calculations did not consider the adsorption of the surfactant at high coverage or at the critical micelle concentration, it can give an

164 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

approximation of its molecular adsorption and its effect towards the borohydride adsorption. It is believed that at low concentrations of Triton X-100, the surfactant molecules are vertically adsorbed on the electrode between the borohydride molecules, as illustrated in Figure 5.13. This avoids the hydrogen atoms, from the borohydride dehydrogenation/hydrolysis, to couple with neighbour hydrogen atoms and evolve hydrogen gas. However, if the concentration of Triton X-100 is increased above a certain value, the catalyst surface might be blocked and could hinder the borohydride oxidation decreasing the current and the fuel efficiencies. Thus, in the optimum concentration, Triton X-100 might be able to increase the selectivity of the catalyst.

BH BH 4 BH H BH 4 3 H 3

Electrode Electrode

Triton H BH H X-100 BH3 BH BH 3 4 Triton X-100 4

Electrode e- Electrode e-

Figure 5.13 Schematic diagram of the breaking up of the surface due to the presence of Triton X-100.

5.2.4 Fuel utilization

During the oxidation process, part of the sodium borohydride fuel reacts while the rest generates hydrogen through the hydrolysis. In order to evaluate the influence of

165 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

surfactants in the amount of sodium borohydride used to generate electricity, the fuel utilization can be calculated via the following equation:

( )

The equation compares the experimental current (Iexp) generated by the direct oxidation of borohydride ions with the theoretical current considering that no hydrolysis occurs.

The term represents the current that would have been generated if the amount of borohydride hydrolysed had been used for the direct oxidation of borohydride, according to reaction (2.1) and can be calculated by Faraday’s law and the amount of borohydride consumed due to its hydrolysis which is related to the hydrogen gas generated. The amount of hydrogen generated was measured from the volume of gas collected in the burette and the borohydride used in hydrogen generation was obtained by stoichiometry from reaction (2.2). The fuel utilization, calculated with and without surfactants at each applied potential during the electrolysis at the Au planar electrode and the Au/C felt electrode, is shown in Figures 5.14 and 5.15, respectively. The data shows that the fuel utilization was higher towards negative potentials and generally higher on the Au/C than on the planar Au electrode.

At potentials between -0.1 and 0.2 V vs. Hg/HgO, in the presence of 0.1 wt. % of SDS and Triton X-100 at the planar Au electrode the fuel utilization slightly decreased between 2 % and 5 % compared to that in the absence of surfactants, due to the increase in the hydrogen generation.

166 ChapterChapter 55:: TheThe effecteffect ofof surfactantssurfactants onon borohydrideborohydride hydrolysishydrolysis andand oxidationoxidation

70 e)

65 d)

Fuel utilization / % % / utilization Fuel 55

c) a) 50 b)

45 -0.4 -0.2 0.0 0.2 0.4 0.6 Electrode potential, E vs. Hg/HgO / V

Figure 5.14 Comparison of the fuel utilization at the planar Au electrode in the presence and in the absence of surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430.

100

80 b) a)

70 c) % / utilization Fuel 60

d) 50 e) -0.4 -0.2 0.0 0.2 0.4 Electrode potential, E vs. Hg/HgO / V

Figure 5.15 Comparison of the fuel utilization at the Au/C felt electrode in the absence and the presence of surfactants: a) No surfactant, b) 0.1 wt. % Triton X-100, c) 0.1 wt. % SDS, d) 0.1 wt. % FC4430 and e) 0.3 wt. % FC4430.

167 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

In the presence of FC4430 and increasing the concentration, the fuel utilization increased by 11 % and 15 % at 0.2 V vs. Hg/HgO and at 0.1 wt. % and 0.3 wt. %

FC4430, respectively. The results suggest that the use of FC4430 might be beneficial for the DBFC when the Au planar electrode is used, as the hydrogen generation rate decreases and the fuel utilization increases. At the Au/C electrode, the presence of SDS slightly increased the fuel utilization, between 2 and 8 %, especially at more positive potentials. However, in the presence of FC4430, the fuel utilization decreased by 18 % and 23 %, at 0.1 wt. % and 0.3 wt. % concentration, respectively. When the planar Au electrode was used, the fuel utilization increased in the presence of FC4430, which inhibited the borohydride hydrolysis and slightly increased the current density.

The different behaviour of this surfactant in the two gold electrodes can be due to the fact that the FC4430 molecules are less likely to be adsorbed on the planar Au electrode than on the Au/C electrode. The gold dispersed nanoparticle electrode has a larger active surface area, providing more available surface for molecules adsorption than that on the planar Au of similar geometrical area. An improved appreciation of surfactant electrosorption awaits future studies on in-situ infra-red and Raman spectroscopies of the electrode surface. The active area ratio of 3-D compared to 2-D surface was estimated from the relation between the current values in the electron transfer controlling region, obtained using the 3-D electrode compared with that obtained using the 2-D electrode. This relation can be obtained from equation (5.6) [186]. The conditions and concentration of borohydride in the two cases was the same, simplifying equation (5.6) to equation (5.7) and substituting the current values, the result suggests that the 3-D active surface is 50 times larger than the 2-D surface.

168 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

( ) ( )

( )

where f (= F/RT), which is a constant equal to 38.92 V-1.

5.3. The effect of surfactants on the kinetic parameters of borohydride oxidation

5.3.1. Planar Au electrode

5.3.1.1. CV at a stationary electrode

Figure 5.16 shows the cyclic voltammetry (CV) of solutions containing 0.02 mol dm-3

-3 NaBH4 in 3 mol dm NaOH (curve a) and in the presence of: 0.1 wt. % SDS (curve b),

0.1 wt. % Triton X-100 (curve c) and 0.1 wt. % FC4430 (curve d), using a planar gold electrode of 0.125 cm2 geometric area. The continuous line (curve a) in Figure 5.16 illustrates the CV of the borohydride solution without surfactants, which oxidation peaks have been previously reported in the literature [32, 33]. The first peak a2, at approximately -0.33 V vs. Hg/HgO and the shoulder a3 up to 0.4 V vs. Hg/HgO are due

- to the direct oxidation of BH4 and the oxidation of intermediate species such as

- BH3OH , respectively. In the reverse scan (not shown for clarity), a third peak c1 corresponds to the oxidation of intermediate species at 0.3 V vs. Hg/HgO [32, 33]. The presence of 0.1 wt. % of SDS increased the oxidation peaks a2, a3 and c1 noticeably, as illustrated in Figure 5.16 (curve b). The CV of a solution containing 0.1 wt. % SDS in 3 mol dm-3 NaOH in the absence of borohydride showed no activity, which demonstrates that the peaks a2, a3 and c1 in curve b) were due to the oxidation of borohydride and its intermediates and that the surfactant is not electroactive in that potential region.

169 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

10 b)

2 - d) 8

/ mA cm mA / a) j j a 6 3

4 Current density, density, Current 2

0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

Electrode potential, E vs. Hg/HgO / V

Figure 5.16 Cyclic voltammogram at a stationary 0.125 cm2 Au disk electrode using a potential sweep rate of 10 mV s-1 in solutions containing 0.02 mol dm-3 -3 0 NaBH4 + 3 mol dm NaOH at 23 C. a) no surfactant, b) with 0.1 wt. % SDS, c) 0.1 wt. % Triton X-100 and d) 0.1 wt. % FC4430.

In the curve c) of Figure 5.16, which corresponds to the solution containing 0.1 wt. %

Triton X-100, the first peak a2, due to the direct oxidation of borohydride ions, decreased approximately 40 % from 9.9 mA cm-2 in the absence of surfactant to 6 mA

-2 cm . The second oxidation wave a3 also decreased when the surfactant was used. The oxidation peak c1, in the reverse scan (not shown), was shifted 0.015 V to more positive potentials and was also inhibited with a decrease in the current density of around 2 mA cm-2 when compared to the CV in the absence of surfactant. No hydrogen oxidation peak was observed, as expected, and the decrease in the current density is attributable to inhibition of borohydride oxidation.

170 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

The CV of a blank solution containing 3 mol dm-3 NaOH and 0.1 wt. % Triton X-100 in the absence of sodium borohydride (not shown) was also performed, indicating that there was no activity with no borohydride in solution and that the oxidation peaks were due to the oxidation of borohydride and its intermediates rather than the oxidation or decomposition of the Triton X-100. Figure 5.16 also illustrates the CV of the borohydride solution containing 0.1 wt. % FC4430 (curve d). The first peak a2, was not observed in this case, which suggests that the surfactant was either inhibiting the direct oxidation of the borohydride ions or shifting it to more positive values. That suggests that more energy is required to drive the oxidation of borohydride due to the overpotential caused by the adsorption of this surfactant on the gold surface. However, what appears to be the second oxidation wave, possibly corresponding to a3 in the absence of surfactants, was obtained at higher current density than that without surfactants.

The reverse scan, shown in Figure 5.17, presented a peak, which was shifted by 0.054 V vs. Hg/HgO to more negative potentials, compared to the oxidation peak c1 with no surfactants; and the oxidation current increased 1.65 mA cm-2 compared to that in the absence of surfactants. The fact that the oxidation of the intermediates also shifts to a more positive values suggests that the surface might be covered by the surfactant and that is causing an activation overpotential.

171 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

20 b)

2 -

/ mA cm mA / j j

d) 8 a)

c) Current density, density, Current 4

0 -0.8 -0.3 0.2 Electrode potential, E vs. Hg/HgO / V

Figure 5.17 Reverse scan of the cyclic voltammogram at a stationary 0.125 cm2 Au disk electrode using a potential sweep rate of 10 mV s-1 in solutions -3 -3 0 containing 0.02 mol dm NaBH4 + 3 mol dm NaOH at 23 C. a) no surfactant, b) with 0.1 wt. % SDS, c) 0.1 wt. % Triton X-100 and d) 0.1 wt. % FC4430.

5.3.1.2. Linear sweep voltammetry at an RDE

The kinetic parameters of the borohydride oxidation, in the absence and presence of surfactants, can be calculated using RDE. The three-electrode cell assembled as shown in the Figure 3.2 using a Pt mesh as the counter electrode, and a saturated calomel electrode (SCE) as the reference electrode. The cathode side was filled with a solution

-3 -3 of 3 mol dm NaOH and in the anode side solutions containing 0.02 mol dm NaBH4, 3 mol dm-3 NaOH, with and without surfactants, were used. A Nafion 117 membrane was used to separate anode and cathode compartments to avoid that the reaction on the counter electrode affects the borohydride oxidation reaction. Cyclic voltammetry was carried out at a potential sweep rate of 10 mV s-1 and rotation rates from 0 to 3600 rpm.

172 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Figure 5.18 shows the cyclic voltammogram at different rotation rates of 0.02 mol dm-3

-3 - NaBH4 in 3 mol dm NaOH at the planar Au RDE. The oxidation of BH4 ions started at -0.6 V vs. Hg/HgO and it was followed by the kinetic and mixed controlled region.

The mass transport controlled region appears between 0.0 V and 0.25 V vs. Hg/HgO, where the limiting current changes proportional to the rotation rate of the electrode. As it can be appreciated in Figure 5.18, the current density drops to zero at about 0.5 V vs.

Hg/HgO. This decay was attributed to the precipitation of oxygenated boron species or boron-oxides formed during the borohydride oxidation or to the presence of H2 bubbles

- formed due to the BH4 hydrolysis at the gold surface, which cover the electrode surface and modify the value of the accessible geometric surface of the electrode [190, 191].

180 3600 rpm

160 2500 rpm

140 1600 rpm

900 rpm

- 120 400 rpm

100

/ mA cm mA /

j 80

40 Current density, density, Current 20

0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 Electrode potential, E vs. Hg/HgO / V

-3 -3 Figure 5.18 Cyclic voltammogram of 0.02 mol dm NaBH4 in 3 mol dm NaOH at 23 oC at the Au RDE (0.125 cm2) rotated between 400 rpm and 3600 rpm.

173 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Since the results in Figure 5.16 showed that the SDS surfactant enhances the oxidation

- of BH4 , different concentrations of this surfactant (0.1 wt. %, 0.01 wt. %, 0.001 wt. %,

0.0001 wt. % and 0.00001 wt. %) were tested to analyse the effect of the surfactants concentration. Figure 5.19 shows the CV of the borohydride oxidation in the absence and in the presence of 0.001 wt. % SDS at controlled rotation rates (400 rpm and 2500 rpm), showing a noticeable increase in the current density.

200 0.001 wt. % SDS 180

160 No surfactant

- 140

120 / mA cm mA /

100 j 80 60 40

Current density, density, Current 20 0 -0.6 -0.4 -0.2 0 0.2 0.4 0.6

Electrode potential, E vs. Hg/HgO / V

Figure 5.19 Cyclic voltammogram at a 0.125 cm2 Au disk electrode at a potential sweep rate of 10 mV s-1 and at 400 rpm and 2500 rpm for 0.02 mol dm-3 -3 NaBH4 + 3 mol dm NaOH in the absence and the presence of 0.001 wt. % SDS at 23 oC.

The CVs in the presence of the other concentrations of SDS tested, showed a similar shape. However, the limiting current plateau between 0 V and 0.4 V vs. Hg/HgO gradually increased with the concentration of SDS from 0.00001 wt. % to 0.001 wt. %

SDS. When the concentration of SDS was further increased to 0.01 wt. % and 0.1 wt.

174 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

%, the limiting current was slightly lower than that in the absence of surfactants. The

Levich plot for solutions in the absence and the presence of the SDS surfactant, can be obtained by representing the limiting current density, from the CV shown in Figures

5.17 and 5.18, versus the square root of the rotation rate. Figure 5.20 shows a typical

Levich plot for solutions in the absence and presence of different concentrations of

SDS.

0.3 No surfactant

0.00001 wt. % SDS

2 - 0.0001 wt. % SDS 0.001 wt. % SDS

0.2 0.01 wt. % SDS

/ mA cm mA / L L j 0.1 wt. % SDS

0.1 Limiting current density current Limiting 0 0 5 10 15 20 Square root of the rotation rate, ω1/2 / rad1/2 s-1/2

Figure 5.20 Limiting current density vs. square root of rotation speed (Levich) plot for borohydride oxidation at a gold RDE at 10 mV s-1. In a solution of -3 -3 0.02 mol dm NaBH4 + 3 mol dm NaOH at room temperature in the absence () and in the presence of 0.00001 wt. % SDS (), 0.0001 wt. % SDS (), 0.001 wt. % () SDS, (◊) 0.01 wt. % SDS and (*) 0.1 wt. % SDS at 23 oC.

The limiting currents, at 0.05 V vs. Hg/HgO, increased approximately by 13 %, 21 % and 37 % in the presence of SDS at concentrations of 0.00001 wt. %, 0.0001 wt. % and

0.001 wt. %, respectively compared to that in the absence of surfactants. The oxidation was mass-transport limited at all rotation rates used in the calculation, as evidenced by

175 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

the linearity in the Levich plot. In the presence of 0.1 wt. % the peak current density due to the borohydride oxidation was 10 % higher than that in the absence of surfactant when the electrode was not rotating (Figure 5.16); but at higher rotation rates, the limiting current was 10 % lower in the presence of 0.1 wt. % SDS, compared with that in the absence of SDS (Figure 5.20).

It has been suggested that the SDS surfactant forms micelles at a concentration of 8 mM

(0.2 wt. %) and that the micelles are adsorbed on a gold electrode in a hemispherical shape [188]. The thickness of the hydrocarbon region in the hemimicellar film is comparable to the length of the hydrocarbon tail of a fully extended SDS monomer.

Therefore, the SDS surfactant is adsorbed in monomers forming monolayers of surfactant of 15 Å thickness (which is the approximate length of the SDS monomer), when the concentration of SDS increases, more monomers are adsorbed forming a monolayer, which is what occurs at concentrations higher than 0.01 wt. %. Although, the concentration of SDS was below the CMC, the effect of the SDS at concentrations higher than 0.01 wt. % have the same effect as the CMC, because either in monolayers or 15 Å thickness (at 0.01 wt. %) or in adsorbed hemimiceles (>0.02 wt. %), the surfactant is partially covering the electrode and hindering the access of borohydride ions [188, 189, 192].

Figure 5.21 shows a comparison of the CV of the borohydride ions at different rotation rates in the absence and in the presence of 0.1 wt. % Triton X-100. For clarity, only the forward scan is presented. In the presence of Triton X-100, the onset potential, potential at which the borohydride oxidation starts, was -0.65 V vs. Hg/HgO, similar to that in the absence of surfactant. However, the oxidation current was generally lower at all

176 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

potentials in the presence of Triton X-100 and two oxidation peaks, are clearly observed, corresponding to two steps during the oxidation reaction of borohydride ions.

180 3600 rpm 160 3000 rpm

140 2500 rpm

2 - 120 1600 rpm 900 rpm 100

/ mA cm mA / Absence of j j 80 Triton X-100

40 Current density, density, Current 20 With Triton X-100

0 -0.6 -0.4 -0.2 0 0.2 0.4 0.6

Electrode Potential, E vs. Hg/HgO / V

- Figure 5.21 Comparison of the CV for BH4 oxidation at an Au RDE (0.125 cm2) in the absence and in the presence of 0.1 wt. % Triton X-100 at 23 oC and different rotation rates.

The reaction rate corresponding to the first step, between -0.2 V and 0.0 V vs. Hg/HgO, seems to be partially controlled by the electron transfer since the current density increased with the potential. However, it was not proportional to the rotation rate of the electrode, as the current density at 900, 1600, 2500 and 3000 rpm remains at values between 20 and 40 mA cm-2. At more positive potentials > 0 V vs. Hg/HgO, the current increased proportionally to the square root of the rotation rate indicating a mass transport controlled process. If the surfactant was adsorbed on the electrode surface inhibiting the initial stages of the borohydride oxidation, this indicates that at more positive potentials the layer of Triton X-100 is probably thinner and the borohydride

177 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

ions can reach the electrode surface. Figure 5.22 illustrates a different oxidation curve obtained from the borohydride oxidation in the presence of 0.1 wt. % FC4430, showing both forward and reverse scans. The CV of a blank solution, containing 0.1 wt. %

FC4430 in 3 mol dm-3 NaOH, was also carried out showing no activity. That suggests that the peaks obtained in the presence of borohydride are not due to the FC4430 decomposition. In the forward scan of Figure 5.22, only one sharp peak at 0.4 V vs.

Hg/HgO was observed. This peak could be due to the delayed oxidation of borohydride ions, which would correspond to the peak a2 in the oxidation of borohydride without surfactants (Figure 5.13).

100 Reverse scan 0 rpm Forward scan 400 rpm

- 900 rpm 1600 rpm

60 2500 rpm

/ mA cm mA /

j 3600 rpm

Current density , , density Current 20

0 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 Electrode potential vs. Hg/HgO / V

Figure 5.22 Cyclic voltammogram at a 0.125 cm2 Au disk electrode using a potential sweep rate of 10 mV s-1 and at different rotation rates for 0.02 mol dm-3 -3 o NaBH4 + 3 mol dm NaOH + 0.1 wt. % FC4430 at 23 C.

It could also be that the FC4430 inhibits the direct oxidation of borohydride and that the peak is due to the oxidation of intermediate species formed during the first step of the

178 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

borohydride hydrolysis. In that case, the peak would coincide with the peak a3 in the oxidation of borohydride without surfactants. In the presence of FC4430, the typical limiting current plateau due to the oxidation current is substituted by a sharp peak. The current density increased with the rotation rate, indicating that the process is mass transport controlled. The reverse scan showed a sharp peak at around 0.2 V vs. Hg/HgO, which corresponds to the oxidation of intermediate species, similar to that corresponding to c1 [32, 33]. Levich plots obtained for the solutions containing 0.1 wt.

% Triton X-100 and 0.1 wt. % FC4430, which also showed complete mass transport controlled oxidation, are shown in Figure 5.23.

0.2 No surfactant

0.1 wt. %Triton 2 - 0.16

0.1 wt. % FC4430

/ mA /cm mA

j 0.12

0.08

0.04 Limiting current density,Limitingcurrent

0 0 5 10 15 20

Square root of the rotation rate, ω1/2 / rad1/2 s-1/2

1/2 -3 Figure 5.23 Comparison of the Levich plot (jL vs. ω ) in a solution of 0.02 mol dm -3 o NaBH4 + 3 mol dm NaOH at 23 C in the absence () and in the presence of 0.1 wt. % Triton X-100 () and 0.1 wt. % FC4430 (). A gold RDE (0.125 cm2) was used at a potential sweep rate of 10 mV s-1.

179 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

The limiting current densities in the presence of 0.1 wt. % Triton X-100 and 0.1 wt. %

FC4430 decreased by an average of 12.4 % and 39 %, at 0.27 V and 0.4 V vs. Hg/HgO, respectively, compared to that in the absence of surfactants. In both cases the highest current was taken as the limiting current. The decrease in the limiting current suggests that these surfactants at 0.1 wt. % hinder the oxidation of borohydride, probably by surface electrode blockage.

- 5.4.1.2.1. Diffusion coefficient of BH4

The Levich equation (5.8) can be used to calculate the diffusion coefficient of borohydride ions, using the limiting current value calculated from similar curves to those in Figure 5.16:

⁄

⁄ ( ) ⁄

where jL is the limiting current, n is the number of electrons changed in the reaction, considered 8 for an Au surface electrode [193], F is the Faraday’s constant (96485 C mol-1), c is the bulk concentration (mol cm-3), ω is the rotation rate (rad s-1), μ is the kinematic viscosity of the solution (cm s-1) and D is the diffusion coefficient (cm2 s-1).

The values of the kinematic viscosity (0.02 cm2 s-1) for the concentration of borohydride used has been taken from the literature [190].

The diffusion coefficient of the borohydride ions in the absence of surfactants was 0.98

-5 2 -1 -3 -3 × 10 cm s for a solution containing 0.02 mol dm NaBH4 in 3 mol dm NaOH at 23 oC. This value can be compared with that reported by Wang et al. [197], who obtained

180 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

an average value of 1.14 × 10-5 cm2 s-1 using the same concentration of borohydride and sodium hydroxide at 30 oC. The difference between these two values, 14 %, can be attributed to the effect of the temperature on the diffusion coefficient. Wang et al. [197]

-3 -3 reported values for borohydride solutions (0.02 mol dm NaBH4 in 2 mol dm NaOH) that varied from 1.09 × 10-5 to 1.62 × 10-5 cm2 s-1 when the temperature changed from

20 to 40 oC. Chatenet et al. [190] obtained a diffusion coefficient of (1.28 0.22) × 10-5

2 -1 -3 -3 cm s for a solution containing 0.01 mol dm NaBH4 in 4 mol dm NaOH, which is also comparable with the values obtained in this work, taking into account that the presence of sodium hydroxide generally increases the viscosity of the solution and thus, decreases the diffusion coefficient [190, 197]. A slightly higher value was reported by

Cheng and Scott [198], 1.68 × 10-5 cm2 s-1, considering a solution containing 1.32 mol

-3 -3 o dm NaBH4 in 2.5 mol dm NaOH at 25 C. The diffusion coefficient calculated by

Cheng and Scott is higher than other values previously reported, probably due to the fact that a higher sodium borohydride concentration was used, however, the authors did not provide any explanation of the difference in the values obtained and those previously reported in the literature.

Table 5.2 summarises the diffusion coefficient of borohydride ions in the presence of

SDS, which increased by 20 % at 0.00001 wt. % SDS, and 30 % and 40 % for 0.0001 wt. % and 0.001 wt. %, respectively compared to that in the absence of surfactants. In the presence of higher concentrations of SDS (0.01 wt. % and 0.1 wt. %) the limiting current density was lower than that in the absence of surfactant, leading to lower values of the diffusion coefficient.

181 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Surfactant level in the Diffusion coefficient of electrolyte borohydride ion, D / 105 cm2 s-1 No surfactant 0.98 0.00001 wt.% SDS 1.21 0.0001 wt.% SDS 1.41 0.001 wt.% SDS 1.63 0.01 wt. % SDS 0.94 0.1 wt. % SDS 0.84 0.1 wt. % Triton X-100 0.77 0.1 wt. % FC4430 0.33

Table 5.2 Diffusion coefficient of borohydride ion in the absence and the presence of various concentrations of SDS, 0.1 wt. % Triton X-100 and 0.1 wt. % FC4430 at 23 oC.

Table 5.2 also includes the diffusion coefficient for a solution of borohydride in the presence of 0.1 wt. % Triton X-100, with a value of 0.77 × 10-5 cm2 s-1, which is 21 % lower than that obtained in the absence of surfactant. The decrease in the diffusion coefficient was due to lower current densities than in the absence of surfactants and may be due to blockage of the electrode surface. The presence of 0.1 wt. % FC4430 also led to a decrease in the diffusion coefficient, with a value of 0.33 × 10-5 cm2 s-1. That value is three times lower than in the absence of surfactants, 0.98 × 10-5 cm2 s-1, which indicates that the presence of FC4430 most likely hinders the access of the borohydride to the electrode surface, consequently minimising the borohydride oxidation and the current density. Another explanation for such a low diffusion coefficient value is that the Levich plot obtained for the solution containing 0.1 wt. % FC4430, shown in Figure

5.23, is slightly shifted from the origin (0, 0), which is a condition that must be fulfilled to use the Levich equation.

182 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

5.4.1.2.2. Heterogeneous rate constant for the oxidation of borohydride

The Koutecky-Levich equation (5.9) can be used to calculate the heterogeneous oxidation reaction rate constant ka under electron and mass transfer mixed control [93]:

⁄ ( ) ⁄ ⁄

Where j is the current density and the other terms have been defined previously. The reciprocal of the current density vs. the inverse of the square root of the rotation rate is a straight line and the ka value can be calculated from the intercept of the curve at infinite rotation rate (equation 5.10), i.e. when ω-1/2 = 0.

( ) ( )

Figure 5.24 shows an example of the above mentioned graph, at different potentials, for

-3 -3 a solution containing 0.001 wt. % SDS, 0.02 mol dm NaBH4 + 3 mol dm NaOH at 23 oC. Table 5.3 shows the rate constant values in the absence of surfactants and in the presence of SDS. The kinetic rate constant increased with the electrode potential for the borohydride oxidation in alkaline solutions in all cases. In the absence of surfactants,

-1 the kinetic rate constant ka at 0 V vs. Hg/HgO for potential sweep rate of 10 mV s was

26 × 10-3 cm s-1. This value is in the same order of magnitude as that obtained by Cheng

-3 -1 -3 and Scott [198] who reported ka = 16 × 10 cm s at 0 V vs. Hg/HgO for 1.32 mol dm

-3 o -1 NaBH4 in 2.5 mol dm NaOH at 25 C using a potential sweep rate of 5 mV s . A

-3 -1 similar value, ka = 12 × 10 cm s , was obtained by Finkelstein et al. [20] at potentials

183 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

between -0.332 V and -0.302 V vs. Hg/HgO, using cyclic voltammetry at 25 mV s-1 in a

-3 -3 o solution containing 0.005 mol dm NaBH4 in 1 mol dm NaOH at 25 C. It was expected that higher kinetic rate constants would be observed at high potential sweep rates, due to the relationship between the size of the diffusion layer on the electrode surface and the potential scan rate used [186], since at a slow potential sweep rate the diffusion layer grows further from the electrode in comparison to a fast scan. At low scan rates, the flux of borohydride ions to the electrode surface is considerably smaller than it is at faster potential sweep rates, resulting in lower currents [186].

- 60

/ mA /

- L

j 50

30 cm

Inverse of the limiting current density, density, current limiting the of Inverse 0 0 0.05 0.1 0.15 0.2 Inverse of the square root of the rotation rate, ω-1/2/ rad-1/2 s1/2

Figure 5.24 Reciprocal current density vs. the inverse of the square root of the rotation rate for a solution containing 0.001 wt. % SDS, 0.02 mol dm-3 -3 o NaBH4 in 3 mol dm NaOH at 23 C and : () -0.25 V vs. Hg/HgO, (*) -0.2 V vs. Hg/HgO, ()-0.15 V vs. Hg/HgO, () 0 V vs. Hg/HgO.

184 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

3 -1 Potential, Rate constant, ka / 10 cm s E vs. No 0.00001 wt. 0.0001 wt. 0.001 wt. 0.01 wt. 0.1 wt. % Hg/HgO surfactant % SDS % SDS % SDS % SDS SDS -0.25 0.51 1.97 2.10 2.82 1.701 3.54 -0.20 0.88 3.47 3.27 4.29 3.374 5.42 -0.15 1.49 6.44 10.80 6.54 6.57 8.56 -0.10 2.35 15.00 13.20 10.4 12.76 13.80 0.00 6.2 26.90 22.30 16.0 18.00 35.11 Table 5.3 Kinetic rate constants for borohydride oxidation at controlled potentials -3 -3 in an electrolyte containing 0.02 mol dm NaBH4 in 3 mol dm NaOH at 23 oC with different concentrations of SDS from 0 wt. % to 0.1 wt. %.

Santos and Sequeira reported two different techniques to calculate the diffusion coefficient of borohydride ion and the kinetic parameters for borohydride oxidation in a

-3 -3 solution of 0.03 mol dm NaBH4 in 2 mol dm NaOH. They initially published the use of chronoamperometric measurements and the Cottrell equation to calculate the diffusion coefficient and to generate the Anson plots (charge per unit area vs. square root of the time) to calculate the kinetic rate constants. Values of the constants were 37

× 10-3 cm s-1 and 36.5 × 10-3 cm s-1 at 0.043 V vs. Hg/HgO and 0.243 V vs. Hg/HgO, respectively [202]. In a second paper, higher values were obtained for the kinetic rate constant at the same potentials, using an Au RDE, 92 × 10-3 cm s-1 and 316 × 10-3 cm s-1

[197]. The authors do not give any explanation of the difference in the values obtained for the kinetic constants, but the difference in the diffusion coefficient values was attributed to the fact that the Cottrell equation is applicable only for diffusion-controlled regions. At low overpotentials, the reaction was partially charge transfer controlled, leading to a not completely diffusion controlled region and larger diffusion coefficients than the previously reported values. The diffusion coefficient values determined from

RDE techniques were considered more accurate than the former reported by Santos et al. [199]. The kinetic rate constants calculated using the Anson plot, were lower than

185 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

the obtained using the RDE technique, which value at 0.243 V vs. Hg/HgO (316 × 10-3 cm s-1) doubles the obtained in this work at the same potential, 147 × 10-3 cm s-1.

Table 5.3 also shows the heterogeneous oxidation rate constant (ka) in the presence of

SDS. The heterogeneous rate constant increased in the presence of SDS in the borohydride solution and also with its concentration. At -0.20 V vs. Hg/HgO, ka increased from 0.9 × 10-3 cm s-1 in the absence of surfactant, to 3.5 × 10-3 cm s-1, 4.3 ×

10-3 cm s-1 and 5.42 × 10-3 cm s-1 in the presence of 0.00001 wt. %, 0.001 wt. % and 0.1 wt. % SDS, respectively. The same trend is followed at the other applied potentials, i.e., the presence of SDS increased the heterogeneous reaction rate constant for borohydride oxidation.

5.4.1.2.3. Standard rate constant for borohydride oxidation

In order to calculate the standard rate constant for the electron-transfer Ks, the log ka vs. potential was plotted. The value of Ks could be determined by the intercept of this plot and a plot of log kc vs. potential, where kc is the rate constant for reduction. The intercept of both lines corresponds to the log Ks and the equilibrium potential. In the case of borohydride, kc cannot be calculated due to the irreversibility of the reaction.

However, the intercept of log ka with the equilibrium potential provides an approximate value of log Ks. The equilibrium potential was taken as the rest potential, which is given by the value at which the current is still zero, immediately before the oxidation onset, which was approximately -0.6 V vs. Hg/HgO as it can be observed in Figure 5.13. By extending the regression line to -0.6 V in the plot log ka vs. potential, shown in Figure

5.25, the value Ks of the reaction can be found.

186 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

Electrode Potential, E vs. Hg/HgO / V -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0

Rest Potential -1

-2

a log k log log Ks -3

-4

-5

Figure 5.25 Log ka vs. Potential for a solution containing 0.001 wt. % SDS, 0.02 mol -3 -3 dm NaBH4 in 3 mol dm NaOH.

-1 The standard rate constant Ks for borohydride oxidation obtained at 10 mV s was 1.62

× 10-5 cm s-1, which is comparable to that obtained by Santos et al. [197] 1.8 × 10-5 cm s-1, using an Au RDE for a sodium borohydride concentration of 0.03 mol dm-3 in 2 mol

-3 dm NaOH. Filkenstein at al. [20] calculated Ks using a solution containing 0.005 mol

-3 -3 dm NaBH4 in 1 mol dm NaOH via cyclic voltammetry at a potential sweep rate of 25

-1 -2 -1 mV s and obtained a value of 6.8 × 10 cm s . The higher values of Ks and ka obtained by Filkenstein at al. [20], compared with that reported by other authors is probably due the use of a higher potential sweep rate, which may have risen to non- steady state currents. In the presence of 0.00001 wt. %, 0.0001 wt. %, 0.001 wt. %, 0.01

-5 wt. % and 0.1 wt. % of SDS, the heterogeneous rate constant Ks increased to 2.16 × 10 cm s-1, 5.1 × 10-5 cm s-1, 1.3 × 10-4 cm s-1, 5.35 × 10-3 cm s-1, 4.71 × 10-2 cm s-1, respectively at 23 oC.

187 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

5.4.1.2.4. Charge transfer coefficient for borohydride oxidation

The charge transfer coefficient, α can be calculated by modifying equation (5.9) for an irreversible system [20, 186]:

( ) ⁄ ⁄ ⁄ [( ) ( ]

Where IL is the limiting current (equation 5.9), Ik is the kinetically limited current, Ks is the standard heterogeneous rate constant, f = F/RT being F Faraday’s constant, R the molar gas constant and T the temperature, E is the potential and Eo is the standard potential. The value of (1-α) can be obtained from the slope of the ln[(I/( IL -I)] vs. E plot represented by equation (5.12) [20]:

( ) ( ) ( )

A value of  = 0.63 was obtained for the borohydride oxidation in the absence of surfactants. The fact that the charge transfer coefficient is larger than 0.5, suggests that the energy barrier of the oxidation reaction is lower than that of the reduction and less overpotential is required for the reaction to obtain larger currents [186, 201]. Santos and

Sequeira [199] observed a moderate variation of α with the borohydride concentration

-3 and with the solution temperature, obtaining a value of 0.84 using 0.03 mol dm NaBH4 in 2 mol dm-3 NaOH at 25 oC. The difference between the values of α obtained by

Santos and Sequeira and that obtained in the present work may be due to the differences in sodium borohydride and sodium hydroxide concentrations. A value of α = 0.67, was

188 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

calculated by Gyenge et al., using the same concentration of borohydride as Santos and

Sequeira and adding 1.5 × 10-3 mol dm-3 TU [32]. They concluded that the presence of

TU hinders the oxidation of borohydride. In the presence of 0.00001 wt. %, 0.0001 wt.

% and 0.001 wt. % SDS, the charge transfer coefficients were 0.58, 0.55 and 0.63, respectively. Comparing these values with that obtained for a solution without surfactant, 0.63, the transfer coefficient decreases in the presence of SDS, making the oxidation reaction more favourable [202].

5.3.2. Au/C electrode

Figure 5.26 shows the comparison of the CVs in the absence and in the presence of 0.1 wt. % SDS at 0 rpm and 2500 rpm (400 rpm to 1600 rpm not shown in Figure 5.23 for clarity). The presence of 0.1 wt. % SDS did not affect the limiting current (IL) compared to that in the absence of surfactants at 0 rpm and 400 rpm. However, from 900 to 2500 rpm, IL was slightly lower in the presence of SDS. At 0 rpm, the oxidation peak was shifted to more negative values in the presence of SDS, -0.44 V vs. Hg/HgO. That shift in the oxidation potential was observed at all the rotation rates, suggesting that the oxidation reaction is slightly favoured by the presence of SDS.

189 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

140

120

2 - 100

/ mA cm mA / 0.1 wt. % SDS

80 , j ,

40 No Surfactant Current density Current

0 -0.8 -0.3 0.2 0.7 Electrode Potential, E vs. Hg/HgO / V

- 2 Figure 5.26 CV of the BH4 oxidation at the Au/C RDE (0.5 cm ) in the absence (black lines) and in the presence (orange lines) of 0.1 wt. % SDS at rotation rates from 0 (continuous lines) to 2500 rpm (broken lines) and at 23 oC.

Figure 5.27 compares the CV at different rotation rates in the absence and in the presence of 0.1 wt. % Triton X-100. It can be observed that the borohydride oxidation potential is shifted by 0.07 V towards negative values in the presence of 0.1 wt. %

Triton X-100 at 0 rpm, compared to that in the absence of surfactants. The oxidation peak potentials seem to be shifted to more negative values, which suggest that less energy is required to start the borohydride oxidation and thus it is favoured by the presence of Triton X-100. However, there is a decrease in the current density when 0.1 wt. % Triton X-100 is in the solution, probably as a consequence of the electrode surface been blocked. As lower concentrations of Triton X-100 showed a decrease in the hydrogen generation rate during the electrolysis at constant current, as previously

190 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

shown in Table 5.1, the CV of a solution containing a lower concentration of Triton X-

100, 0.001 wt. %, was analysed and it is presented in Figure 5.28.

140

120

2 -

100 / mA cm mA /

No surfactant

j 80

Current density, density, Current 0.1 wt. % Triton X-100

0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6

Electrode Potential, E vs. Hg/HgO / V

-3 - -3 Figure 5.27 Cyclic voltammogram of 0.02 mol dm BH4 in 3 mol dm NaOH at 23 oC at the Au/C RDE (0.5 cm2) in the absence (black lines) and in the presence (orange lines) of 0.1 wt. % Triton X-100 at rotation rates of 0 rpm (continuous lines) and 2500 rpm (broken lines).

In this case, the potential at which the oxidation reaction starts was also slightly shifted to more negative potentials but the current density did not decrease compared to that in the absence of surfactants at low rotation rates (0-900 rpm) and at larger rotation rates

(900-2500 rpm), the current density slightly increased as it can be observed in Figure

5.28. The current density peak increased 13.4 % at 3000 rpm in the presence of 0.001 wt. % Triton X-100. This suggests that Triton X-100, at 0.001 wt. %, inhibited the borohydride hydrolysis, leading to a higher amount of free molecules of borohydride that can be oxidised and increase the current density. These results agree with those

191 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

obtained in the previous section and indicate that there must be an optimum concentration of Triton X-100, between 0.001 wt. % and 0.01 wt. %, which inhibits the hydrogen generation rate and does not affect negatively to the current density.

Therefore, as it has been reasoned in the previous section (5.2.3), the presence of Triton

X-100 at the optimum concentration (below the CMC) might be beneficial to the DBFC by decreasing the hydrogen generation rate and not affecting negatively to the oxidation of borohydride and the current density.

140

120

- 100 No surfactant

/ mA cm mA /

j j

40 0.001 wt. % Triton X-100 Current density, Current 20

0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6

Electrode Potential, E vs. Hg/HgO / V

- 2 Figure 5.28 Cyclic voltammogram of the BH4 oxidation at the Au/C RDE (0.5 cm ) in the absence (black lines) and in the presence (red lines) of 0.001 wt. % Triton X-100 at rotation rates at 0 rpm (continuous lines) and 2500 rpm (broken lines) at 23 oC.

192 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

-3 -3 The CVs of 0.02 mol dm NaBH4 in 3 mol dm NaOH in the absence (black lines) and in the presence of 0.1 wt. % FC4430 (blue lines) and 0.3 wt. % FC4430 (green lines) are presented in Figure 5.29. The presence of 0.1 wt. % and 0.3 wt. % of FC4430 decreased the current density at the Au/C electrode. No value of CMC data has been reported for this surfactant until date, and maybe 0.1 wt. % is already higher concentration than the CMC. In that case, the micelles can hinder the access of the borohydride ions to the electrode, decreasing the current density. The electrolysis reported in the previous section showed that the presence of FC4430 also decreases substantially the hydrogen generation rate, by an average of 15 % and 45 % for concentrations of 0.1 wt. % and 0.3 wt. %, respectively.

120

100

2 -

No surfactant

/ mA cm mA /

j 60 0.3 wt. % FC4430

0.1 wt. % FC4430

40 Current density, density, Current 20

0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6

Electrode Potential, E vs. Hg/HgO / V

- 2 Figure 5.29 Cyclic voltammogram for BH4 oxidation at a Au/C RDE (0.5 cm ) at rotation rates of 0 rpm (continuous lines) and 2500 rpm (broken lines) in the absence (black lines) and in the presence of 0.1 wt. % FC4430 (blue lines) and 0.3 wt. % FC4430 (green lines) at 23 oC.

193 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

However, the decrease in current density, nearly 60 % at 0.1 wt. % FC4430, compared to that in the absence of surfactants, makes it inappropriate to be used for the DBFC. It is possible that at concentrations of FC4430 lower than the CMC, the current density is not affected by the surfactant, which would be similar behaviour to that observed with the other non-ionic surfactant tested, Triton X-100.

5.4. Conclusions

Surfactant addition to the electrolyte plays an important role in the direct and indirect borohydride fuel cells, as the additive can affect the hydrogen generation rate and the partial current density for borohydride oxidation. At the planar Au electrode, the use of the non-ionic surfactant Triton X-100 increased the hydrogen generation rate, and decreased the fuel utilization and the current density. That suggests that the use of

Triton X-100 with the planar Au electrode favours the hydrolysis reaction rather than inhibit it. However, promising results were obtained at the Au/C electrode when 0.001 wt. % Triton X-100 was added to the borohydride alkaline solution. The current density from the borohydride oxidation was unaffected by the presence of Triton X-100, while the hydrogen generation rate decreased by 23 %. DFT studies showed that the presence of Triton X-100, decreases the free energy of adsorption of the borohydride ions and thus, it favours the borohydride oxidation. At low coverage, the Triton X-100 molecule would be vertically adsorbed on the Au(111) surface, between borohydride molecules, hindering the H-H bonding that leads to hydrogen evolution. At the optimum concentration (lower than the CMC), Triton X-100 could minimize the borohydride hydrolysis in DBFC with an Au/C anode electrode, as it seem to decrease the gas generation rate, increasing the fuel efficiency and maintaining the power density. On the contrary, the anionic fluorosurfactant S-228M seems to favour the hydrolysis reaction,

194 Chapter 5: The effect of surfactants on borohydride hydrolysis and oxidation

increasing the hydrogen generation rate at both planar Au and Au/C electrodes. That might be useful in an IBFC, where the objective is to generate hydrogen from sodium borohydride. Due to its anionic nature, S-228M competes with the borohydride ions for the occupation of the active sites on the catalyst surface, reducing the surface availability for borohydride oxidation and the current density.

At the planar Au electrode, the hydrogen generation rate increased by 25 % in the presence of 0.1 wt. % SDS, increasing also the current density. The diffusion coefficient of borohydride and the kinetic constants for borohydride oxidation also increased with the concentration of SDS. These results suggest that SDS favours both, the borohydride oxidation and its hydrolysis reaction rates. On the contrary, at the Au/C electrode SDS has the opposite effect and slows the incomplete oxidation. In the presence of 0.1 wt. %

SDS, the hydrogen generation rate and the current density slightly decreased compared to that with no surfactants [202]. It is believed that increasing the concentration of SDS to a value closer to and below the CMC (0.2 wt. % [203, 204]), will further decrease in the hydrogen generation.

The presence of low concentrations of Zonyl FSO (a nonionic fluorocarbon surfactant) increased the hydrogen generation rate at both planar Au and Au/C electrodes, decreasing the fuel efficiency and the experiments suggest that this type of surfactant is not effective for the reduction of borohydride hydrolysis. The non-ionic fluorinated surfactant FC4430 inhibited hydrogen generation at both planar Au and Au/C electrodes and it substantially decreased the current density. It is assumed that the surfactant

FC4430 blocks the active sites of the electrode, hindering both borohydride oxidation and its hydrolysis.

195 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

The majority of studies on the direct oxidation of borohydride ions have been reported using microstructured electrodes surfaces where the gold catalyst or other electroactive material is supported on carbon powder, or on solid two-dimensional electrodes. This approach limits the space-time yield of the fuel cell which could be improved by the use of three-dimensional porous electrodes [205]. Typically, three-dimensional electrode materials used in electrochemistry include reticulated structures such as copper [206], nickel [207], aluminium [54], and reticulated vitreous carbon (RVC) [205]. Due to its low cost, high conductivity, light weight and wide variety of porosity, RVC has been extendedly used in electrochemical processes such as; metal recovery [191], hydrogen peroxide production for the removal of organic compounds [208] and energy conversion in batteries and fuel cells [209]. The use of three-dimensional electrodes for the direct oxidation of borohydride is a novel approach that could potentially maximise the amount of borohydride ions being oxidised per unit volume in a fuel cell and consequently improve the power output from the anode.

Physical vapour deposition (PVD) gold coated RVC could be an appropriate three- dimensional electrode for borohydride oxidation, due to the low density, low thermal expansion, high corrosion resistance and high thermal and electrical conductivities. It has a high surface area, rigid structure, low resistance to fluid flow, and very high resistance to temperatures in non-oxidizing environments [205]. Reticulated vitreous carbon electrodes, with porosity between 10 ppi and 100 ppi, were gold coated during 1 min, 2 min and 3 min using the PVD technique, obtaining a coating thickness in the

196 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

order of nanometres. Figure 6.1 shows the SEM image of an RVC electrode of 80 ppi, gold-coated for a deposition time of 3 minutes, which presents an approximate coating thickness of approximately 200 nm, as shown in the figure inset.

100 nm

100 m

Figure 6. 1 SEM image of a gold-coated RVC 80 ppi prepared by a sputtering technique over 3 min, the inset shows the thickness of the deposit on the carbon substrate [109].

6.1. Cyclic voltammetry

In order to analyse the effectiveness of these electrodes as anode material for the borohydride oxidation, CV was carried out on each gold coated RVC electrode. Figure

6.2 shows the CVs performed with solutions containing 0.02 mol dm-3 (0.075 wt. %)

-3 NaBH4 in 3 mol dm (12 wt. %) NaOH using a 10 ppi RVC 1 min PVD gold coated electrode as the anode material and potential sweep rates from 20 mV s-1 to 200 mV s-1.

Figure 6.2 also shows the CVs carried out on a solid gold electrode (0.125 cm2 geometric area) at the same electrolyte conditions and at 20 mV s-1 potential sweep rate.

For comparison, the figure illustrates the CV of a solution containing only 3 mol dm-3

(12 wt. %) NaOH and no borohydride, and the CV of a solution containing 0.02 mol

197 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

-3 -3 dm (0.075 wt. %) NaBH4 in 3 mol dm (12 wt. %) NaOH on an uncoated RVC electrode, both at 150 mV s-1. No current was observed when the uncoated RVC electrode was tested in the presence of borohydride and the gold coated electrode was tested in the absence of borohydride, which confirms that the bare RVC is not active towards the oxidation of borohydride and that the electrochemical activity of the gold- coated RVC is due to the gold nanoparticles. For simplicity, only the anodic scan is shown in the figure. The shape of the cyclic voltammogram on the solid gold disc electrode exhibits the characteristic peaks a2 and a3 that have been previously reported in the literature [210]; the first one a2, at - 0.470 V vs. SCE (-0.33 V vs. Hg/HgO) due to the direct oxidation borohydride followed by a wide oxidation wave a3, extending up to

+ 0.42 V vs. SCE possibly due to the oxidation of reaction intermediates.

The curves for the oxidation of borohydride on the 10 ppi gold-coated RVC electrode show two oxidation processes, denoted by two oxidation peaks, a2’ and a3’, at 0 V and

0.4 V vs. SCE, respectively, obtained in the CV at 20 mV s-1 using the 10 ppi RVC gold coated (1 min). These peaks are possibly due to the same processes observed on the solid gold electrode a2 and a3, direct oxidation of borohydride ions and oxidation of intermediate species, respectively. The potential of the oxidation peak a2’ appeared shifted 0.470 V to more positive potentials compared with the same peak on the gold solid electrode, meaning that higher activation overpotentials (between 200 – 300 mV) and thus higher energy was required to oxidise the borohydride on the three- dimensional electrodes.

198 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

−3 −3 Figure 6. 2 Cyclic voltammograms of 0.02 mol dm NaBH4 in 3 mol dm NaOH on a 0.125 cm2 gold disk electrode and on a 10 ppi RVC gold-coated electrode (1 × 0.5 × 0.5 cm) at coating time of one minute at different potential sweep rates and 298 K.

However, this can be compensated by the higher current densities (per geometric area) provided at the three-dimensional electrode. The potential of the peak a2’ was moved with the potential sweep rate from approximately 0 V vs. SCE at 20 mV s-1 to 0.3 V vs.

SCE at 200 mV s-1. This potential change with the sweep rate is characteristic of the irreversible nature of the oxidation of borohydride and the increase of IR drop typically observed in porous electrodes. The oxidation of the possible intermediates, a3’, appeared at similar potentials in the gold-coated RVC and in the solid gold electrode,

0.25 V and 0.4 V vs. SCE, respectively. The oxidation peak a3’ was also slightly shifted to more positive values and eventually overlapped by peak a2’ in a wide peak as the

199 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

potential sweep rate increases. During the reverse sweep potential, the curves also presented the sharp oxidation peak c1’, (not shown in the figure for simplicity) assigned

− to the oxidation of some intermediate products such as BH3OH , as reported in the literature [210].

Similar current vs. potential curves as those shown in Figure 6.2 were also obtained for the other gold-coated RVC ppi grade electrodes (20, 30, 45, 60, 80 and 100 ppi) at 2 min and 3 min gold deposition times showing similar features: three oxidation peaks a2’, a3’ and c1’ which were shifted towards positive values when the potential sweep rate increased.

The charge transfer coefficient can be calculated from the slope of the equation for irreversible systems, assuming an electrochemical single step for the oxidation of borohydride; the following equation can be used [200]:

⁄ ⁄ { } { ( ) [( ) ] } (6.1) ( )

0’ where E' is the formal electrode potential for the oxidation of borohydride, na is equal to 1, the number of electrons involved in the rate determining step, D is the diffusion coefficient of the borohydride,  the charge transfer coefficient, ν is the potential sweep rate, Ks is the heterogeneous rate constant and f (= F/RT), which is a constant equal to

38.92 V-1.

200 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

3 30 ppi 2.5 10 ppi

60 ppi

2 80 ppi

SCE / V SCE /

1.5 Ep vs. Ep

0.5 Peak Potential, Peak

-0.5 -5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 Logarithm of the potential sweep rate, ln v / V s-1

Figure 6. 3 Potential at the peak current vs. natural logarithm of the potential sweep rate for 1 minute PVD sputtering gold deposited RVC electrodes.

Figure 6.3 shows the plots of Ep vs. ln(v) used to calculate the charge transfer coefficient for the gold coated RVC electrodes using equation (6.1). A charge transfer coefficient α, between 0.87 and 0.98 was obtained, depending on the porosity grade

(ppi) of the RVC electrode used. Those values are slightly larger than the value reported by Santos et al. [200], 0.84, under similar conditions on a solid gold electrode. This suggests that the gold coated RVC electrodes show better or at least similar catalytic activity than the solid gold electrode. In order to calculate the number of electrons transferred through the borohydride ions oxidation reaction, the slope of the peak current vs. the square root of the potential sweep rate for gold-coated RVC ppi grades obtained after 1, 2 or 3 minutes deposition time was calculated. The curves in Figures

201 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

6.4, 6.5 and 6.6, show peak current densities that increased with the porosity of the electrode for the same deposition time and it also increased with the deposition time for the same porosity. That could be due to an increase of the space-time yield for higher porosity materials. The figures also show that the peak current of the oxidation processes increased with the square root of the potential sweep rate. However, not all the RVC electrodes exhibited linear behaviour, which can be attributed to a non- uniform gold deposition or some defects on the RVC surface structure that hindered the adherence of gold on the carbon substrate. Assuming that the oxidation of borohydride ions occur in a single-step irreversible process, the following expression can be used to correlate the relationship between the peak current and the potential sweep rate [201]:

⁄ ⁄ ⁄ ( )[( ) ] (6.2)

Where Ip is the peak current, A is the geometrical surface area of the RVC electrodes, n is the total number of electrons and c is the concentration of borohydride ions. The diffusion coefficient of borohydride ions calculated from various electrochemical techniques at 298 K are: 2.1 × 10-5 cm2 s-1 from polarography studies, 1.6 × 10-5 cm2 s-1 using Au microelectrodes [211] and varies with temperature from 1.09 ± 0.05 × 10-5 cm2 s-1 at 293 K to 2.36 ± 0.07 × 105 cm2 s-1 at 333 K, calculated via chronoamperometry on a spherical Au electrode [197].

202 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

0.25 100 ppi 80 ppi 0.2 60 ppi 20 ppi

10 ppi / A /

0.15

p I

0.1 Peak current, current, Peak 0.05

0 0 0.1 0.2 0.3 0.4 0.5 1/2 -1 1/2 (Potential sweep rate, ν) / (V s ) Figure 6. 4 Peak current vs. potential sweep rate for 1 minute PVD sputtered gold coated RVC electrodes.

0.25

45 ppi 0.2 20 ppi

10 ppi

/ A / 0.15

p I

0.1

Peak current, current, Peak 0.05

0 0 0.1 0.2 0.3 0.4 0.5 (Potential sweep rate, ν)1/2 / (V s-1)1/2

Figure 6. 5 Peak current vs. potential sweep rate for 2 minutes PVD sputtering gold deposited RVC electrodes.

203 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

0.25 60 ppi 40 ppi 0.2 30 ppi

20 ppi

10 ppi

/ A / 0.15

p I

0.1 Peak current, current, Peak 0.05

0 0 0.1 0.2 0.3 0.4 0.5 1/2 -1 1/2 (Potential sweep rate, ν) / (V s ) Figure 6. 6 Peak current vs. potential sweep rate for 3 minutes PVD sputtering gold deposited RVC electrodes.

For the calculation of the number of electrons, an average of the diffusion coefficient was taken (2.1 × 10-5 cm2 s-1) [197, 212]. The charge transfer coefficient was taken from the previous calculations which gave a value of α between 0.87 and 0.98, depending on the porosity grade (linear pores per inch, ppi) of the RVC electrode used. The RVC geometric area was calculated, taking into account the porosity level, from a linear plot of the surface area per unit RVC volume (considering an electrode volume of 0.25 cm3), versus the porosity. The plot of the specific surface area for different porosity grades 10,

20, 50, 60, 80 and 100 ppi was obtained from a publication by Friedrich et al. [205], who obtained the surface areas of RVCs electrodes of different porosities from SEM images, by measuring the struts and circle areas in a defined region of the RVC sample.

204 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

Although it has been found that gold does not necessarily yield the full 8 electrons predicted, this value can be used as an approximation; for example, Santos et al. [197] suggested effectively 7.6 electrons transfer at 298 K and using a concentration of borohydride of 0.03 mol dm-3 (0.11 wt. %). Other authors have reported different values depending on the oxidation potential and suggested that the number of electrons released seems to be influenced by the further oxidation of some of the intermediates

[32, 33, 198]. In this work, the number of electrons transferred during the oxidation of borohydride on the different gold-coated RVC electrodes using equation (6.2) varies from 5.5 to 10.3. Although the average value was 8 electrons, the reason for obtaining a value higher than 8 suggests that there are superimposed currents due to further oxidation reactions of some intermediates, uneven potential and current distribution on the three dimensional electrodes, together with the fact that a homogeneously uniform gold coating was considered for the calculation of the geometric area of the RVC electrodes, which can cause the net number of electrons to be larger than the predicted 8

[198].

6.2. Kinetic constants

The heterogeneous rate constant of the oxidation process Ks can be calculated using the following equation:

1 na F  o  I  0.227nFAc  K exp E  E (6.3) p BH s   p  4  RT  

205 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

The values of the kinetic rate constants ka were also calculated using equation 6.4 and are presented in Table 6.1 together with the values of the constants obtained for different gold-coated RVC electrodes.

( ) {[ ] ( )} ( )

The values of the heterogeneous rate constant are in the order of 10-5 cm s-1 and can be

-5 -1 compared with those obtained by Santos et al. [213], Ks = 1.8 × 10 cm s using cyclic voltammetry on a solid gold electrode with an area of 3.14 × 10-2 cm-2 and a concentration of 0.03 mol dm-3 sodium borohydride in 2 mol dm-3 NaOH. The values obtained using the gold coated RVC electrodes are generally higher than that reported by Santos et al., the difference can be attributed to the larger area of the three- dimensional electrodes. However, the values obtained in this work are two orders of magnitude lower when compared with those reported by other authors using different

-2 -1 techniques. Finkelstein et al. [210] reported a much larger value of Ks, 6.8 × 10 cm s , which was obtained from the cyclic voltammetry of a solution of 0.005 mol dm-3

-3 -1 NaBH4 in 1 mol dm NaOH at a potential sweep rate of 25 mV s and using Koutecký-

Levich equation.

The difference in the Ks values between the different authors could be due to the choice of the E0 value, -1.48 V vs. SCE [193] and the use of a larger sweep rate. As previously discussed in Chapter 5, it is expected that due to the relationship between the diffusion layers thickness on the electrode surface and the potential scan rate used, higher kinetic constants are obtained at high potential sweep rates [186], since at a slow potential

206 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

sweep rate the diffusion layer grows further from the electrode in comparison to a fast scan. At low scan rates, the flux of borohydride ions to the electrode surface is considerably smaller than it is at faster potential sweep rates, resulting in lower currents

[186].

Potential vs. SCE / V Porosity PVD -0.5 -0.3 -0.1 0.1 0.3 0.5 grade of Ks sputtering RVC 5 3 time / / 10 Kinetic rate constant at different potentials, ka / 10 / ppi min cm s-1 cm s-1

1 15 7 15 30 70 150 330 10 2 7 10 30 70 200 540 1500 3 25 7 15 25 50 100 200 2 15 6 15 30 60 130 270 20 3 40 5 9 15 25 40 70 1 70 4 6 9 15 20 30 30 2 60 2 3 4 6 7 10 3 75 5 7 10 15 20 30 45 3 50 8 15 25 40 70 130 1 50 1 1 1 2 2 2 60 3 180 7 10 15 20 20 30 80 1 40 1 1 1 1 2 2 1 20 0.5 1 1 1 1 1 100 3 50 1 1 1 1 1 2

Table 6. 1 Heterogeneous and kinetic rate constants for the different gold coated RVC electrodes at 298 K. The concentration of borohydride ions was -3 -3 0.02 mol dm NaBH4 in 3 mol dm NaOH.

As it can be seen in Table 6.1, and with some exceptions, the values of the heterogeneous rate constant Ks tend to increase slightly with the porosity grade and the

207 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

deposition time. The values of the kinetic rate constant ka, reported in the Table increase with the electrode potential and decrease with the porosity grade of the electrode. At less positive potentials the difference is not significant but as the electrode potential becomes more positive, the difference increases, this result suggests that on lower ppi gold-coated RVC electrodes such as 10, 20 and 30 the reaction is faster than on the higher grades.

6.3. Conclusions

The results of cyclic voltammetry at static electrodes and linear sweep voltammetry the rotating disc electrode PVD gold deposited RVC electrodes showed that three- dimensional electrodes are advantageous for the oxidation of borohydride ions. The current due to the oxidation of borohydride increases when the porosity of the gold- coated RVC electrodes increases from 10 to 100 ppi at constant deposition times.

Similar situation occurs when the gold deposition time increases for the same porosity grade gold-coated RVC electrodes; the general trend is that the current due to the oxidation of borohydride increases as larger amount of gold is deposited. The charge transfer coefficient varied between 0.87 and 0.98 depending on the ppi and the deposition time on the RVC electrodes. The average number of calculated electrons transferred during the borohydride oxidation was 8, although in some cases lower (5.5) and larger (10.3) number of electrons were found. This is probably due to the uneven current and potential distribution on the three dimensional electrodes and that the current measured might be a sum of the current generated from the borohydride oxidation and the oxidation of hydrogen from the borohydride hydrolysis. The heterogeneous and kinetic rate constants of borohydride on these electrodes tended to increase slightly with the porosity grade and the deposition time, having the same order

208 Chapter 6: PVD gold coated RVC electrodes for borohydride oxidation

of magnitude, 10-5 cm s-1 as those reported in the literature using cyclic voltammetry.

However, when compared with values obtained using rotating disc electrodes, the values are smaller, probably due to the difference in the surface area and the election of the E0 value, -1.48 V vs. SCE [193], to calculate the kinetic constants.

209 Chapter 7: Hydrogen generation from sodium borohydride

Chapter 7: Hydrogen generation from sodium borohydride

The most promising technology for transport and niche applications, the Polymer

Electrolyte Membrane (PEM) Fuel Cell has remained as a practical demonstrator and applications beyond buses (e.g., Ballard [194]) are now beginning to emerge such as cars (e.g., the Honda Clarity [195]), motorcycles (e.g., the Intelligent Energy demonstrator [196]). Higher power and higher charge capacity fuel cell devices could emerge as an alternative sustainable energy if fuel cell technology and hydrogen supply infrastructure can be implemented. The supply, storage and safe handling of hydrogen gas, however, are still problematic due to its highly explosive nature and the difficulty in storing it at high pressures [4]. An alternative approach is the use of hydrogen- containing compounds such as LiBH4, NaBH4 or KBH4 which can be stored safely in concentrated aqueous solution and deliver hydrogen on demand [18]. Among the advantages of using sodium borohydride as a hydrogen source for a fuel cell, the following are important: it can be stored safely as a powder form or as a 30 wt. % concentrate in strong alkaline solutions at pH 14, with a half-life time of approximately

430 days [4], sodium borohydride contains 10.6 wt. % of hydrogen which of course will be lower in a 30 wt. % aqueous solution [19] and the oxidation product of borohydride, metaborate, is environmentally acceptable and can, potentially, be recycled to sodium borohydride [22].

Hydrogen stored in borohydride can be released in a separate reactor through the hydrolysis of borohydride [14, 15] according to reaction (1.2). This reaction is favoured when the electrolyte is not strongly alkaline and in the presence of catalysts such as Pt,

Pd, Ru, Ni, NixB, Co, CoyB (where y = 1 or 2) or Pt-LiCoO2 among others which are less active [139, 214]. Catalyst degradation is a very well-known difficulty. The

210 Chapter 7: Hydrogen generation from sodium borohydride

agglomeration of catalyst particles, surface oxidation or partial dissolution in alkaline

NaBH4 solutions, are common reasons for catalyst deterioration [15]. At a high hydrogen production rate, the strong generation of hydrogen bubbles in the liquid phase promotes collision between catalyst particles that can contribute to damage the material

[138].

In this chapter, the results of the hydrogen generation using a two compartment reactor are reported using nitrogen gas to pressurize the fuel chamber. A solution containing 4 mol dm-3 (15. % wt) of sodium borohydride is forced into a compartment that holds a catalyst, using low pressure nitrogen gas. The components are contained in the catalytic compartment and their concentration at any given time changes as the reaction is in a batch mode. The pressure within the catalytic compartment increases due to the hydrogen production and the volumetric flow rate is measured. As the reaction proceeds, the reactant and the catalyst are in contact, producing metaborate and hydrogen gas. If the concentration of metaborates significantly increases, inhibition or blocking of active catalyst sites can result in an overall decrease in the rate of hydrogen generation. Catalysts such as Pd-Ir alloy, Pt nanoparticles on carbon paper and Pd deposited on granular carbon (Pd/C) were evaluated together with the effect of the catalyst loading and the importance of the quality of the water used to prepare the solution.

7.1. Generation of hydrogen using platinised titanium catalyst

The platinised titanium catalyst consisted of a cylindrical titanium mesh of 70 mm diameter and 60 mm height coated with platinum (ca. 3.5 m). Two different solutions

211 Chapter 7: Hydrogen generation from sodium borohydride

of sodium borohydride in sodium hydroxide were used in the hydrogen generator

-3 -3 -3 reactor to test the catalyst: 0.02 mol dm NaBH4 in 2 mol dm NaOH and 4 mol dm

-3 NaBH4 in 0.5 mol dm NaOH. In both cases, the solutions were prepared with distilled water. A negligible amount of hydrogen generation was obtained during both series of experiments. Platinised titanium is an excellent catalyst for the hydrogen evolution, particularly during the electrolysis of water to produce oxygen and hydrogen [215, 216], when it operated as an anode for borohydride oxidation without the assistance of anodic current, platinised titanium appeared inactive under the experimental conditions.

7.2 Generation of hydrogen using Pd-Ir catalyst supported on carbon fibres

The catalyst, consisting of Pd-Ir nanoparticles deposited on carbon microfibers and supported on a titanium foil substrate plate, has already been used as a cathode material for a DBFC [86]. The microfibrous carbon catalyst support was fabricated by Patrissi et al. [165] by applying a 30-100 kV pulse between a carbon plate, which contained carbon microfibres (0.5 mm length and 10 m diameter), and the titanium foil substrate plate, which contained a carbon adhesive film. Around 125,000 fibres per cm2 were deposited on the titanium foil. A catalyst loading of 12.3 mg cm-2, with a composition of 50:50 Pd:Ir, was obtained by linearly cycling the electrode containing the carbon

-3 -3 -3 -3 microfibres with a solution containing 2 × 10 mol dm PdCl2 and 2 × 10 mol dm

-3 -3 o Na2IrCl6H2O in 0.2 mol dm KCl and 0.1 mol dm HCl at 70 C. The cyclic voltammetry was carried out at 1 mV s-1 between -0.15 V and -0.30 V vs. Ag/AgCl

[165]. Figure 7.1 shows the hydrogen generation rate versus time when two 32 cm2 plates of containing the Pd-Ir catalyst on one side were used. The solutions containing 4

212 Chapter 7: Hydrogen generation from sodium borohydride

-3 -3 mol dm NaBH4 (15 wt. %) NaBH4 in 0.5 mol dm (2 wt. %) NaOH were prepared using different types of water; distilled water, tap water and river water. When distilled

3 -1 3 -1 -1 water was used, an average hydrogen generation of 80 cm min (100 cm min gmetal ) was obtained for approximately 200 minutes. After that time, the rate of H2 generated was insufficient to be accurately measured.

200

-1 150

min

/ cm

100

Hydrogen generation rate, rate, generation Hydrogen

0 0 50 100 150 200 Time, t / min

-3 Figure 7. 1 Comparison of the hydrogen generation from 4 mol dm NaBH4 in 0.5 mol dm-3 NaOH using a Pd/Ir catalyst using:  tap water,  river water and ■ distilled water to prepare the solution.

When tap water was used, the hydrogen generation rate rapidly decreased from 200 cm3

-1 3 -1 -1 3 -1 3 -1 -1 min (260 cm min gmetal ) at the start, to less than 10 cm min (13 cm min gmetal ) after 10 minutes. With river water, the flow rate decreased from 180 cm3 min-1 (230 cm3

-1 -1 3 -1 3 -1 -1 min gmetal ) to 30 cm min (39 cm min gmetal ) after 40 minutes. The poor hydrogen

213 Chapter 7: Hydrogen generation from sodium borohydride

generation rates using tap and river water suggest that the catalyst could have been poisoned by dissolved contaminants contained in the water. The typical components of tap water include chlorine, fluorine compounds, trihalomethanes, nitrates, pesticides and salts of aluminium, copper, lead, mercury, cadmium and barium [217]. The experimental results suggest that one or more of these components inhibit the decomposition of borohydride.

7.3 Generation of hydrogen using nanoparticulate platinum on carbon

Figure 7.2 shows the hydrogen generation rate versus time when a commercial catalyst consisting on platinum nanoparticles supported on carbon paper (Pt/C) with loading of 4 mg cm2 was used. Two plates having a total geometrical area of 70 cm2 were introduced in the catalyst holder of the reactor. The experiment was carried out with 350 cm3 of a

-3 -3 solution containing 4 mol dm (15 wt. %) NaBH4 in 0.5 mol dm (2 wt. %) NaOH prepared with distilled water. The hydrogen generation rate was 170 cm3 min-1 (600 cm3

-1 -1 min gmetal ) initially, decreasing gradually during the first hour to remain constant at

3 -1 3 -1 -1 30 cm min (100 cm min gmetal ) during the second hour. This electrode disintegrated during operation, leaving catalyst particles in the solution which hindered the control over the reaction due to the impossibility to separate catalyst and solution during operation. The disintegration of this catalyst suggested that Pt/C catalyst supported on carbon paper was not stable enough to be considered for the generation of hydrogen from borohydride.

214 Chapter 7: Hydrogen generation from sodium borohydride

160

140

-1

min 120

/ cm

v 100

Hydrogen generation, generation, Hydrogen 40

0 0 20 40 60 80 100 120 140 160 180 200

Time, t / min

-3 -3 Figure 7. 2 H2 generation from 4 mol dm (15 wt. %) NaBH4 in 0.5 mol dm (2 wt. %) NaOH using nanoparticulate platinum supported on Vulcan carbon printed on Toray carbon paper. The solution was prepared with distilled water. Initial temperature: 23 oC (296 K). Final temperature: 32 oC (305 K). No measurements were taken during the experiment on this occasion.

7.4. Palladium 0.5 wt. % supported on granular carbon catalyst

(Pd/C)

7.4.1. Importance of the water source in the hydrogen generation: distilled water, drinking water and river water

Three different solutions containing 4 mol dm-3 (15 wt. %) of sodium borohydride, initially without NaOH were prepared with distilled water, commercial still drinking water and river water. The catalyst, 10 g Pd 0.5 wt. % supported on granular carbon,

215 Chapter 7: Hydrogen generation from sodium borohydride

was loaded in the reaction chamber of the glass reactor previously described in the section 3.6. A new charge of catalyst was used in each experiment.

The still drinking water showed high activity towards the hydrolysis of borohydride and hydrogen gas was generated vigorously while the solution was prepared in the absence of sodium hydroxide. The hydrogen generation was such that it was impossible to introduce the solution into the lower part of the reactor. To slow down the hydrogen generation and be able to pour the solution into the reactor 0.5 mol dm-3 (2 wt. %)

NaOH was added to the solution. Figure 7.3 shows that the flow rate of hydrogen in this

3 -1 3 3 -1 -1 experiment was around 300 cm min (6 × 10 cm min gmetal ) during the first 10 minutes falling to 140 cm3 min-1 after 20 minutes and 30 cm3 min-1 just before the 60 minutes of the experiment were achieved. The pressure decreased from 0.8 bar to 0.2 bar (0.8 × 10-5 N m-2 to 0.2 × 10-5 N m-2) following a similar decreasing trend to the flow rate, as seen in Figure 7.3.

Despite that the hydrogen was generated, while the solution was being prepared with river water and no stabilizer (i.e. no NaOH), it was possible to introduce the solution inside the lower compartment of the reactor using this type of water. As soon as the solution was forced to the upper part of the reactor to be in contact the catalyst, vigorous hydrogen gas generation was observed which caused, in a few seconds, an increase in the pressure to 1.0 bar (1 × 10-5 N m-2) as seen in Figure 7.4.

Similarly the flow rate increased gradually during the first 15 minutes reaching 400 cm3

-1 3 3 -1 -1 min (8 × 10 cm min gmetal ). The flow rate decreased gradually from this point

3 -1 3 3 -1 -1 down to 110 cm min (2.8 × 10 cm min gmetal ) at 30 minutes. The pressure also

216 Chapter 7: Hydrogen generation from sodium borohydride

decreased from 1 to 0.8 bar and remained at 0.8 bar (0.8 × 10-5 N m-2) until the end of the experiment. When distilled water was used to prepare the solution, the average

3 -1 3 3 -1 -1 hydrogen flow rate was 400 cm min (8 × 10 cm min gmetal ) over 100 minutes then

3 -1 3 3 -1 -1 the flow rate remained constant at 300 cm min (6 × 10 cm min gmetal ) during the

3 -1 3 3 -1 -1 next 60 minutes, to gradually decrease to 50 cm min (50 × 10 cm min gmetal ) after

300 minutes of operation.

600

500

-1

min

3 400

/ cm

300

200

Hydrogen generation rate, rate, generation Hydrogen 100

0 0 50 100 150 200 250 300

Time, t / min

3 Figure 7. 3 H2 flow rate when 10 g of Pd/C catalyst were used with 350 cm of -3 -3 solutions containing 4 mol dm NaBH4 in 0.5 mol dm NaOH prepared with:  still drinking water, and × river water. And the 350 cm3 of -3 solution containing 4 mol dm NaBH4 and no NaOH prepared with:  distilled water.

217 Chapter 7: Hydrogen generation from sodium borohydride

1.6

1.4

1.2

1.0

/ bar

p 0.8

Presure, 0.6

0.4

0.2

0.0 0 20 40 60 80 100 120 140

Time, t / min

Figure 7. 4 Pressure measured during H2 generation when 10 g of Pd/C were used 3 -3 -3 with solutions of 350 cm containing 4 mol dm NaBH4 in 0.5 mol dm NaOH and prepared with:  still drinking water and × river water. And 3 -3 the 350 cm of solution containing 4 mol dm NaBH4 and no NaOH prepared with:  distilled water.

An appreciable amount of hydrogen was generated using river water. The average generation rate, however, was just above half of the hydrogen generated using distilled water, as illustrated in Figure 7.3. The use of still drinking water showed a vigorous hydrogen generation rate in the absence of NaOH, however hydrogen generation was comparable to that when the solution was prepared with river water during the first 60 minutes. The experiments demonstrate that the quality of the water strongly affects the hydrogen generation rate, obtaining the largest hydrogen generation rate and pressure using distilled water to prepare the solutions.

218 Chapter 7: Hydrogen generation from sodium borohydride

7.4.2. Effect of catalyst loading

In order to investigate if the hydrogen flow rate could be increased by using more catalyst, experiments with 10 g, 15 g and 20 g loadings of palladium on granulated carbon (Pd/C), were performed. Solutions of 350 cm3 containing 4 mol dm-3 (15 wt. %)

NaBH4 without NaOH were prepared with distilled water. The curves in Figures 7.5 and

7.6 show the development of the flow rate and pressure respectively, in the system during these experiments.

In the presence of 10 g catalyst, the hydrogen generated varied between 300 cm3 min-1 and 400 cm3 min-1 during 180 min and then gradually decreased for the next 120 min until 10 cm3 min-1. In order to maintain the flow rate constant, the addition of borohydride solution to the catalytic upper chamber of the reactor was controlled by manually open/closing the 3-way valve, which occasioned variation in pressure and hydrogen generation rates in Figures 7.5 and 7.6, specially between 30 and 120 min with 10 g of catalyst used. When 15 g and 20 g of catalyst were used, the hydrogen

3 -1 3 3 -1 -1 generation rate was 600 cm min (8 × 10 cm min gmetal using 15 g Pd/C and 6 ×

3 3 -1 -1 10 cm min gmetal in the presence of 20 g Pd/C) from the time when the solution first made contact with the catalyst. The solution was gradually added to the reaction chamber (every 15-20 min) until all the solution (350 cm3) was in the reaction chamber, after approximately 100 min. The flow rate remained constant during ≈ 120 minutes, followed by a decrease in the hydrogen generation in both cases. This decay of the flow rate was more gradual when 15 g of catalyst was used suggesting that this is the optimum amount of catalyst and that in order to obtain a larger amount of hydrogen, the stabilizer, sodium hydroxide, should not be used. However, problems of hydrogen

219 Chapter 7: Hydrogen generation from sodium borohydride

generation arise during the preparation of the solution as borohydride tends to hydrolyze in a solution without sodium hydroxide.

3 Figure 7. 5 Measurement of the H2 flow rate from 350 cm of a solution containing 4 -3 mol dm NaBH4 and without NaOH in distilled water and different weights of catalyst.  10 g Pd/C  15 g Pd/C and  20 g Pd/C.

Table 7.1 compares the results obtained in this work with the literature for Pt and Pd based catalysts. One of the highest hydrogen generation rates was reported by Peña-

Alonso et al. [135] at 90 dm3 min-1 g-1, however this is an extrapolated value from the generation of 8 cm3 of accumulative hydrogen in 20 minutes. The authors appear to use a very low amount (41-47 mg) of a very active catalyst (Pt and Pd on carbon nanotubes) and perhaps the extrapolation is unrealistic. Certainly, this assumption would need to be demonstrated in a scaled-up reactor. In this work, when the Pd on granular carbon was

3 -1 -1 used, a constant rate of 8 dm min g metal was measured as the highest hydrogen

220 Chapter 7: Hydrogen generation from sodium borohydride

generation rate reported during the longest period of time using a solution of sodium borohydride and a Pd based catalyst. Bai et al. [141] and Wu et al. [140] used the same amount of Pt/C catalysts but with different loadings 13.1 wt. % and 20 wt. %, respectively and the same concentrations of NaOH and NaBH4. Their results show a

3 -1 -1 3 -1 -1 small difference, 23 dm min g metal and 22 dm min g metal , respectively, corresponding the highest hydrogen generation rate to the lowest catalyst loading used.

3.0

2.5

2.0

/ bar

p 1.5

Pressure, 1.0

0.5

0.0 0 50 100 150 200 250 300 Time, t / min

Figure 7. 6 Pressure measured during H2 generation when different catalyst loading:  10 g Pd/C  15 g Pd/C and  20 g Pd/C and 350 cm3 of a solution -3 containing 4 mol dm NaBH4 without NaOH prepared with distilled were used.

221 Chapter 7: Hydrogen generation from sodium borohydride

H2 Operation Catalyst Amount H2 generation NaOH NaBH4 Reactor/ Ref generated time of rate (wt. %) (wt. %) solution 3 3 (cm ) (min) catalyst (dm /min/g met) (mg) 6500 180 Pt/C 28 0.13 2 15 Batch Present reactor: work 350 cm3 16000 200 Pd-Ir/C 700 0.1 2 15 Batch Present reactor: work 350 cm3 59500 170 Pd/C 50 6-8 2 15 Batch Present reactor: work 350 cm3 72000 120 Pd/C 75 8 0 15 Batch Present reactor: work 350 cm3 8 20 Pt and Pd 4.1-4.7 90 0.4 0.1 Fuel flow [135] on carbon not nanotubes specified 1.1 3 13.1 100 23 5 10 Fuel [141] wt. % flow: 10 Pt/C cm3 /min 45 20 20 wt. % 100 22 5 10 Fuel flow [140] Pt/C not specified 1100 45 20 wt. % 0.025 5 5 5 Batch [218] Pt/C reactor: 10 cm3 500 70 1 wt. % 0.25 2.9 0.04 1.5 Batch [219] Pt/C reactor: 25 cm3

Table 7. 1 Comparison of the results obtained in this work and relevant literature.

222 Chapter 7: Hydrogen generation from sodium borohydride

This suggests two possibilities: that better results (higher hydrogen generation rates) could be obtained with a lower catalyst loading, or that scaling up the experimental results leads to a higher value that in reality might not be obtained. Bai et al. obtained

1.1 cm3 in 3 min extrapolating this result to a hydrogen generation rate of 23 dm3 min-1

-1 3 g metal ; at higher scale, Wu et al. obtained 45 cm in 20 min which, using the same

3 -1 -1 extrapolation, resulted on a lower hydrogen generation rate, 22 dm min g metal . The second possibility can justify the low hydrogen generation rate obtained in this work

3 -1 -1 (0.13 dm min g metal ) for the Pt/C catalyst, as the experiment was carried out at a higher scale were 6,500 cm3 were generated in 180 min.

7.4.3. Effect of NaOH stabilizer on the hydrogen generation

A comparison of the generation of hydrogen in the absence and in the presence of 0.5 mol dm-3 (2 % wt) of stabilizer NaOH and using 15 g of catalyst, is shown in Figure

7.7. In the absence of the stabilizer, the flow rate was maintained slightly over 600 cm3

-1 3 3 -1 -1 min (8 × 10 cm min gmetal ) during 100 minutes before decreasing gradually to 100

3 -1 3 3 -1 -1 cm min (1.3 × 10 cm min gmetal ) at 180 minutes. The arrows in the curve indicate the times at which the 3-way valve was open to allow borohydride solution into the upper compartment of the reactor. The solution was introduced into the reaction chamber at a temperature of 25 oC (298 K) and it increased up to 78 oC (351 K) during the first 55 minutes of reaction, to gradually decrease to 29 oC (302 K) during the following 125 minutes that the reaction lasted. In contrast, the hydrogen flow rate when

3 -1 3 -1 -1 NaOH was used barely reached 60 cm min (80 cm min gmetal ) while the pressure was maintained below 0.8 bar. The presence of the stabilizer NaOH decreases the hydrogen generation rate substantially even if a Pd/C catalyst was used and suggests that NaOH should not be used if hydrogen is required very quickly to power a fuel cell.

223 Chapter 7: Hydrogen generation from sodium borohydride

Figure 7. 7 Flow rate of H2 and pressure when 15 g Pd/C catalyst were used with 350 3 -3 cm of a solution containing 4 mol dm NaBH4 in the absence of NaOH (, ) and in the presence of 0.5 mol dm-3 NaOH (, ) all prepared with distilled water. The dark grey arrows show the time at with the solution was added into the upper part of the reactor.

7.4.4. Effect of the reused catalyst

Figure 7.8 shows the comparison of the hydrogen flow rate when 20 g Pd/C catalyst was used consecutively up to three times. The first time the catalyst was used (black

3 -1 3 3 -1 -1 circle symbol) the flow rate was maintained at 600 cm min (6 × 10 cm min gmetal )

3 -1 3 3 -1 -1 over 50 minutes decreasing to 500 cm min (5 × 10 cm min gmetal ) until 100 minutes after the reaction started. Meanwhile the pressure varied between 1 – 2 bar

(shown in Figure 7.9) up to 50 minutes and then nearly dropped to 0 bar for the rest of the experiment.

224 Chapter 7: Hydrogen generation from sodium borohydride

Figure 7. 8 Comparison of the flow rate of H2 when 20 g Pd/C was used three times:  1st use,  2nd use and  3rd use. In all the experiments 350 cm3 of deionised water without stabiliser was used to prepare 4 mol dm-3 NaBH4.

The reason of the large variation in pressure was due to the manual control of the flow

meter valve in order to maintain the flow rate at the maximum level. When the catalyst

was used for the second time (white circle symbol in Figures 7.8 and 7.9) the flow rate

3 -1 3 3 -1 -1 was maintained at over 600 cm min (6 × 10 cm min gmetal ) for nearly 110 minutes

while the pressure increased gradually up to 3 bar at 45 minutes and remained at this

value until nearly 100 minutes. The performance of the catalyst the third time that it

was used (black triangle symbol in Figures 7.8 and 7.9) was not as good as in the

previous ones; this time the flow rate was maintained at 600 cm3 min-1 during the first

225 Chapter 7: Hydrogen generation from sodium borohydride

3 -1 3 3 -1 -1 45 minutes and gradually dropped to 400 cm min (4 × 10 cm min gmetal ) and then to 100 cm3 min-1 at 80 minutes time.

3.0

2.5

2.0

/ bar

p 1.5

Pressure, 1.0

0.5

0.0 0 20 40 60 80 100 120 140 160

Time, t / min

st nd Figure 7. 9 H2 pressure when 20 g Pd/C was used three times:  1 use,  2 use and  3rd use. In all the experiments 350 cm3 of deionised water without -3 stabiliser was used to prepare 4 mol dm (15 wt. %) NaBH4.

The pressure was also high at 2.8 bar but dropped gradually towards zero after nearly one hour of operation. It is worth noting that, in these experiments, the distilled water used to prepare the borohydride solution was cooled down to ≈ 3 oC in order to reduce the spontaneous hydrolysis reaction during the preparation and allow more time to pour the solution into the lower chamber of the reactor without substantial decomposition of the borohydride solution.

In theory, the amount of hydrogen that should be generated from a 350 cm3 solution of borohydride at 4 mol dm3 (15 wt. %) concentration, according to the borohydride

226 Chapter 7: Hydrogen generation from sodium borohydride

3 hydrolysis reaction (1.2), is 125,500 cm . The percentage of H2 generated compared to the theoretical amount was 62 % when the catalyst was used for the first time, 55 % for the second time and 33 % when it was used a third time. This suggests that the catalyst becomes passive by the presence of borates which could be deposited on the palladium particles during the experiments. However, it should be pointed out that when the catalyst had been used; the initial hydrogen generation rate was maintained at 600 cm3

-1 3 3 -1 -1 min (6 × 10 cm min gmetal ) but the total amount of hydrogen generated decreases with the number of times the catalyst was used. Simagina et al. [220] reported rapid activity loses of the Pd/C catalyst and that the hydrogen generation rate approached the same hydrogen generation rate in the presence and the absence of the catalyst. They attributed the disintegration of the catalyst to the high capacity of Pd/C to adsorb hydrogen. In this work, we did not observe such a severe disintegration and after several hours, the catalyst was still active enough to notice the difference between the hydrogen generation rate in its presence and its absence. Besides Simagina et al. [220] explanation, it is likely that some catalyst is lost in the rinse waters used to clean it, as the granulated carbon seems to disintegrate into very small particles during the operation. Another possibility is the presence of borates on the catalyst surface [150,

221-226] but this needs further corroboration.

7.4.4 Hydrogen generation: mix of solid NaBH4 powder and Pd/C catalyst followed by the addition of still drinking water

In order to investigate if the hydrogen generation could be produced using a dry mixture, 53 g of sodium borohydride and 10 g of Pd/C catalyst were mixed together and placed inside the upper compartment of the reactor without NaOH. The lower chamber of the reactor was filled with 350 cm3 of still drinking water and was pressurized to 1

227 Chapter 7: Hydrogen generation from sodium borohydride

bar in order to allow the transfer of water into the upper section of the reactor. Figure

7.10 shows the hydrogen generation rate and the pressure vs. time measured during this experiment. When all the water was added into the upper part of the reactor, the flow meter was closed during the first 30 minutes while the pressure built up to 0.6 bar (0.6 ×

10-5 N m-2). The valve of the flow meter was opened to maintain the pressure at 0.6 bar

(0.6 × 10-5 N m-2) while measuring the corresponding hydrogen flow rate, which was constantly increasing at approximately 4.7 cm3 min-1.

600 0.8

500

-1 0.6

min 3 400

/ cm

/ bar

p 300 0.4

Pressure, 200

0.2

Hydrogen generation, generation, Hydrogen

100

0 0.0 0 20 40 60 80 100 120 140 160 180

Time, t / min

Figure 7. 10 Flow rate of H2  and pressure  when a mixture of sodium borohydride and Pd/C catalyst were mixed together at the upper part of the reactor and still drinking water was gradually added.

The hydrogen flow rate continuously increased over a 2 hours and 15 minutes period,

3 -1 3 3 -1 -1 when it reached 520 cm min (10.4 × 10 cm min gmetal ). After this, the humidity condensed in the flow meter made it difficult to take further measurements and the

228 Chapter 7: Hydrogen generation from sodium borohydride

experiment was manually stopped before the reaction had completed. Although this experiment could have lasted longer if the silica reservoir had been larger to dry the gas produced, it still demonstrates that storing sodium borohydride and a catalyst as a dry mixture, could be an option to generating hydrogen by simply adding water.

7.4.5. Scanning electron microscopy (SEM) imaging of unused and used Pd/C catalyst

In order to evaluate the state of the catalyst before and after the operation, SEM images of the three times used Pd/C catalyst were taken and compared with images of the unused catalyst. The catalyst was granular carbon loaded with 0.5 % wt. of metallic palladium, with an average size of 0.3 – 0.6 cm length and 0.1 – 0.2 cm thickness. The details of the porous structure of the granulated carbon can be clearly appreciated in the main scanning electron microscopy (SEM) image of Figure 7.11.a) that shows porous of

10 – 30 m width. The EDX analysis of the squared area in Figure 7.11.a) shows an average weight percentage of palladium of 40 wt. % as shown in Table 7.2. As expected, the other main elements are carbon and boron. The inset in Figure 7.11.a) shows a closer image of one of the porous walls where the palladium particles were appreciated on the edge with a depth of the pore that appears to be of several μm. This image shows that the palladium particles are spheroidal and granular with an approximate diameter of 50 nm homogeneously distributed in the area.

The EDX spectra analysis of the area inside the circled section in the inset SEM image of Figure 7.11.b) shows that, as expected, the catalyst contains a high concentration of palladium per unit weight. Table 7.3 below shows that the percentage of palladium in

229 Chapter 7: Hydrogen generation from sodium borohydride

this point is 69.05 wt. % which is larger than the average palladium content in a larger area, 40 wt. %, shown in Table 7.2.

Element % Weight Atomic% Boron 8.6 15.1 Carbon 48.5 76.2 Silicon 1.2 0.8 Potassium 1.7 0.8 Palladium 40.0 7.1

Table 7. 2 Elemental composition of the area within the square section of an unused Pd/C catalyst shown in Figure 7.11.a).

This is reasonable as the analysis of the inset SEM in Figure 7.11.b) concentrates around a palladium particle whereas the area in Figure 7.11.a) is larger and includes the carbon support. Figures 7.11.c) and 7.11.d) show the SEM image of a section of the

Pd/C catalyst after being used three times. The average EDX spectra analysis of the area within the square on the main image in Figure 7.11.c) is shown in Table 7.4 and indicates an average content of 10.43 wt. % of palladium with the majority being 77.69 wt. % carbon. Comparing this value with the one shown in Table 7.2 (for a similar surface area) where the percentage of palladium was 40.04 wt. %, indicates that 75 % of palladium has been lost after the catalyst has being used three times. It is interesting to note the presence of sodium, although in small quantities, this was not observed in the

SEM analysis of the unused Pd/C catalyst. The weight percentage of boron remained similar to the unused catalyst. The inset in Figure 7.11.c) shows a close up of the structure of the used catalyst where the porosity of the carbon can still be appreciated.

230 Chapter 7: Hydrogen generation from sodium borohydride

Figure 7. 11 SEM images of: a) the unused Pd/C. The inset shows a higher magnification of the centre of the main image. b) the palladium particles deposited on one of the pore wall of the unused granulated carbon catalyst.c) the three times used Pd/C catalyst. The inset shows an enhancement image from the centre of the main picture. d) the locations of the EDX analysis of the Pd/C catalyst which had been used three times.

Element % Weight Atomic% Boron 9.1 25.5 Carbon 21.8 54.9 Palladium 69.0 19.6

Table 7. 3 Elemental composition of the circle area in the inset SEM image of Figure 7.11.b) of an unused Pd/C catalyst.

231 Chapter 7: Hydrogen generation from sodium borohydride

Element % Weight Atomic % Boron 9.6 11.7

Carbon 77.7 85.6 Na 2.3 1.3

Palladium 10.4 1.3

Table 7. 4 Elemental composition within the large square area of the main SEM image of Figure 7.11.c) of a section of the three times used palladium catalyst on granulated carbon.

A more detailed analysis of the percentage weight of palladium using the SEM image shows that it has been decreased substantially as a result of having previously been used three times in the generation of hydrogen. The cross marks shown on Figure 7.11.d) represent the points were an EDX spectra analysis was carried out and Table 7.5 summarizes the content of each element on these points. If this data is compared with the weight percentage of palladium found in the unused catalyst of 69.05 %, shown in

Table 7.2, with 6.05 wt. % and 6. 15 wt. % in the areas C and D respectively of the image in Figure 7.11.b), the loss of palladium can be appreciated as the used catalyst contains less than 10 % of the initial amount of palladium. The comparison is only a guide because the image of the unused palladium in Figure 7.11.b) was carried out at a magnification of 50,000 times whereas the magnification in Figure 7.11.d) was 10,000 times. However this supports the fact that palladium gradually detaches from the granulated carbon support.

232 Chapter 7: Hydrogen generation from sodium borohydride

Element % Weight Atomic % A B C D A B C D Boron 12.1 14.4 13.4 13.1 13.8 16.7 15.7 14.6

Carbon 82.8 77.9 78.1 84.7 84.7 81.6 82.3 84.8

Sodium 2.2 1.7 2.3 0.9 1.2 0.9 1.3 0.5

Palladium 2.8 6.0 6.1 1.3 0.3 0.7 0.7 0.1

Table 7. 5 Percentage weight of each element found on the cross marks points shown on Figure 7.11.d) using EDX spectra analysis.

7.5. Nitrogen adsorption on the Pd/C catalyst

Figure 7.12 shows the isotherm of nitrogen adsorption on the surface of the Pd/C catalyst.

340

320

-1

3 0.0014

/cm 0.0012

300 0.0010

0.0008

- 1]

[

v 0.0006

Specific volume, Specific volume, 280 0.0004

0.0002

0.0000 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

p/p0 260 0.0 0.2 0.4 0.6 0.8 1.0 Normalised pressure, p/p 0

Figure 7. 12 Nitrogen adsorption isotherm of the surface of the Pd/C catalyst at 77 K  adsorption and  desorption. Inset: Linear transform of adsorption isotherm for determination of BET surface area.

233 Chapter 7: Hydrogen generation from sodium borohydride

The black symbols represent adsorption whereas the white ones depict the desorption process. According to IUPAC protocols [227], the isotherm follows an H1 type curve, indicating the microporous nature of the sample; the small hysteresis observed at 0.4 < p/p0 < 0.95 indicates the presence of mesopores.

In order to calculate Brunauer, Emmett and Teller (BET) surface area, the data from the

N2 adsorption isotherm for relative pressure between 0.1 and 0.3 were transformed into linear form (shown in the inset of Figure 7.12). The curve shows a slope equal to 0.0048 g cm-3 with an intercept of -1.542 × 10-4 g cm-3, which corresponds to a 933.9 m2 g-1 of

BET specific surface area of catalyst.

Figure 7.13 shows the Barrett-Joyner-Halenda (BJH) pore size distribution of the Pd/C catalyst determined from both adsorption and desorption branches of the N2 adsorption isotherm. Both curves showed monotonic growth of the pore volume when the pore size decreased. The peak on the desorption curve at 3 nm < p/p0 < 4 nm can be attributed to a technical artifact associated with the collapse of the film of liquid nitrogen during the desorption process. The sample shows a significant content of micropores and presence of mesopores. The total specific pore volume of the sample is 0.52 cm3 g-1 most of which is attributed to the micropores. These data are consistent with the morphology of the catalyst determined by electron microscopy.

Indeed, according to Figure 7.11.b) and 7.11.d), the catalyst support consists of carbon nanoparticles of ca. 100 nm size with spheroidal morphology. The majority of micropores are situated inside of these carbon nanoparticles whereas mesopores are formed between them during their agglomeration into the larger solid structure. Despite

234 Chapter 7: Hydrogen generation from sodium borohydride

the transport of reactants in the liquid phase being inside the pores, the microporous carbon is usually considered as a rate limiting stage due to the hindered diffusion inside the micropores, the small size of nanoparticles and the hierarchical structure of material

(featuring the mesopores, which can accelerate diffusion of dissolved NaBH4 to the catalyst deposited on the surface of carbon), improve overall performance of catalyst for hydrogen generation.

0.06

0.05

-1

min 0.04

-1

d 0.03

0.02

Pore volume d Pore volume

0.01

0.00 2 3 4 5 6 8 10 20 30 40 50 Log of the average pore diameter, d / nm

Figure 7. 13 BJH pore size distributions of Pd/C catalyst determined from both:  adsorption and  desorption branches in the N2 adsorption isotherm.

7.6. Conclusions

A hydrogen generator was constructed and its performance was evaluated using catalysts including palladium iridium alloy, platinum nanoparticles on carbon paper and palladium deposited on granular carbon. The following conclusions could be reached:

235 Chapter 7: Hydrogen generation from sodium borohydride

1) Palladium on granulated carbon was the most effective catalyst among the

materials tested, by the contact of 15 g of catalyst and a solution of 4 mol dm-3

NaBH4 (15 wt. %) prepared with distilled water and without NaOH stabilizer in

the reaction chamber, a constant hydrogen generation rate of 600 cm3 min-1 (8 ×

3 3 -1 -1 10 cm min gmetal ) was obtained during 120 minutes. That corresponds to a

space-time yield of 0.149 m3 m-3 h-1 or 6.741 mol m-3 h-1, considering the

volume of the whole reactor.

2) Among the different readily available water sources, distilled water was the

one with which optimum results were obtained. However, high hydrolysis

activity was observed with the use of commercial drinking water, and further

experiments should be performed in order to be able to measure the hydrogen

generation rate obtained from this water without any stabilizer. This could be

achieved by cooling the water before preparing the solution in order to

temporarily decrease the hydrogen generation until the solution was poured into

the reactor. The hydrogen generation rate measured during experiments carried

out with river water of tap water was appreciable but insufficient for the fuel cell

applications, probably due to the presence of impurities. Further experiments are

needed to determine the species which poisons the catalyst, i.e. studies of

individual contaminants and their concentration level.

3) A constant increase in the hydrogen generation rate of approximately 4.7 cm3

min-1 was achieved during at least 170 min reaching over 500 cm3 min-1 (10 ×

3 3 -1 -1 10 cm min gmetal ), by mixing the sodium borohydride with the catalyst in the

reaction chamber and adding 350 cm3 still drinking water from the fuel

236 Chapter 7: Hydrogen generation from sodium borohydride

chamber. This experiment might be useful as a different procedure to store

sodium borohydride to produce hydrogen on demand for a H2/O2 fuel cell a

sudden and constant hydrogen generation rate must be provided.

3) Although Pd on granular carbon seems to be a good catalyst for the borohydride

hydrolysis, the loss of palladium during the hydrogen evolution is an issue to

consider. EDX analysis suggested that 75 % palladium was lost after three

operations, probably due to the poor mechanical properties of carbon. A more

robust material such as nickel or stainless steel should be used in order to firmly

support the catalyst over long operations and greater number of cycles.

237 Chapter 8: Conlusions and Further Work

Chapter 8: Conclusions and Further Work

The main conclusions of this work are described, together with a summary of the most important problems that still need to be overcome for the effective use of both direct and indirect borohydride fuel cells.

The use of sodium borohydride for fuel cells offers a promising alternative to incumbent electrical power generation technologies (primarily based on fossil fuels), for semi-large-scale applications (i.e. remote or backup power), as well as small-scale applications such as laptops or mobile phones. Sodium borohydride is safe, easy to handle and it can be used to generate hydrogen by mixing it with water (hydrolysis reaction) or it can be directly oxidised, and act as a fuel for the direct borohydride fuel cell. The hydrolysis of borohydride can be beneficial if the objective is to generate hydrogen for a H2/O2 fuel cell (IBFC) or it can be one of the most important drawbacks if the objective is to use it in the direct oxidation of borohydride. The competition

- between BH4 direct oxidation reaction and its hydrolysis is a function of the electrode material, electrolyte composition and operation conditions such as temperature, concentration of sodium borohydride and sodium hydroxide.

8.1. Direct borohydride fuel cells

DBFC are potentially sustainable, clean, safe and efficient sources of energy. However, several challenges must be addressed to increase the energy density, which remains far below from the theoretically predicted value. Two of the most important drawbacks that have been studied in this work are the search for a more appropriate anode material for the borohydride oxidation, which minimises the hydrolysis reaction and the effect of the

238 Chapter 8: Conlusions and Further Work

addition of surfactants to the electrolyte solution in order to minimise the hydrolysis reaction.

8.1.1. Anode materials

In this work, anode materials such as Pd-Ir alloy deposited on microfibrous carbon and gold coated RVC electrodes have been tested with the aim to compare their performance and to find an appropriate electrode for borohydride oxidation, which is catalytic towards the borohydride oxidation but does not promote the hydrolysis.

8.1.1.1. Pd-Ir coated microfibrous carbon for borohydride oxidation

Pd-Ir coated microfibrous carbon presents low hydrogen generation rates and higher currents than the Au flat plate or the Au/C electrodes during the oxidation of borohydride at the same concentration. However, the Pd-Ir alloy is expensive and causes the spontaneous generation of hydrogen at the open circuit potential (-0.99 V vs.

Hg/HgO). At potentials from -0.8 to 0 V vs. Hg/HgO, the hydrogen generation rate

3 -1 -3 decreased to less than 0.1 cm min in solutions containing 0.5 mol dm BH4 in 3 mol dm-3 NaOH.

DFT studies were carried out at the Pd2Ir1 and Pd2Ir2 surfaces in order to analyse the borohydride oxidation mechanism and to evaluate the effect of the Ir concentration. The reaction mechanism on both Pd-Ir surfaces showed a favourable 8 electron release, according to the relative energies of the intermediate species. However, the activation energies became too high in the last elementary steps, which involve B-OH binding.

The activation barriers were generally lower on Pd2I2 than on Pd2Ir1, as the reaction mechanism was more favourable in the former. From BOH at Pd2Ir1 and BOHOH at

239 Chapter 8: Conlusions and Further Work

Pd2Ir2, the activation barriers became too high and the prediction is that the reaction stops with the release of 5 and 6 electrons at Pd2Ir1 and Pd2Ir2, respectively. The intermediate species, BOH at Pd2Ir1 and BOHOH at Pd2Ir2, might cover the catalyst surfaces reducing its activity towards the borohydride oxidation.

It is possible that the DFT approach used for the calculation of the activation barriers of the B-OH forming reactions, gives energy barrier values that are too high, due to the inaccurate calculation caused by the non-consideration of solvation. Computational models that include the addition of two H2O molecules associated to the BH4 molecule could be utilized. If this reduces the activation energy barriers corresponding to B-OH bond formation, the final oxidation product, metaborate, could be reached, releasing 8e-.

A comparison of the borohydride adsorption energies on Pd(111) and the Pd-Ir alloys showed that, by adding Ir to pure Pd, the borohydride oxidation is favoured. DFT studies were also able to predict the experimental results: at high hydrogen coverage -

1.0 V vs. Hg/HgO, there is hydrogen generation but at more positive potentials the oxidation of borohydride is more favourable than the hydrolysis. The Pd-Ir alloy seems to be a promising catalyst for borohydride oxidation but its design should be improved using DFT predictions to avoid hydrogen generation at the OCV. Different Pd-Ir alloys configurations in the presence and the absence of a third element should be studied.

8.1.1.2. RVC gold coated electrode

PVD gold coated RVC electrodes of different porosities were tested as anodes and it was shown that the current density increased with the porosity from 10 to 100 ppi when they were gold coated at the same period of time. The current density also increased with the gold deposition time due to a larger area being covered on the RVC surface.

240 Chapter 8: Conlusions and Further Work

The calculated charge transfer coefficient on the gold-coated RVC electrodes varied between 0.87 and 0.98 depending on the ppi and the deposition time. The number of electrons transferred varied between 5.5 and 10.3 possibly due to the uneven current and potential distribution on the three dimensional electrodes but the average was 8 electrons. The heterogeneous and kinetic rate constants for the oxidation of borohydride ion were comparable with those reported in the literature using cyclic voltammetry; however, the values were smaller than those obtained at the RDE.

8.1.2. Effect of surfactants in the borohydride oxidation and its hydrolysis

The addition of surfactants to the alkaline borohydride electrolyte alters the hydrogen generation rate due to the borohydride hydrolysis and affects the kinetic parameters of the borohydride oxidation. The effect of surfactants on the borohydride oxidation and its hydrolysis is complex and depends on the nature of the surfactant, its concentration and the electrocatalyst used. The measurement of the hydrogen generated during the electrolysis at different constant currents and constant potentials on a planar Au electrode and on an Au nanoparticulate electrode surface, using solutions in the absence and the presence of surfactants including Zonyl FSO, S-228M, Triton X-100, SDS and

FC4430, permitted the study of the effect of each surfactant on each electrode material.

Rotating disc electrode techniques were used to analyse the effect of the same surfactants on the diffusion coefficient and the kinetics of the borohydride oxidation.

The results showed that the addition of surfactants to the borohydride alkaline solution could increase or decrease the hydrogen generation rate and affect the kinetic parameters of the borohydride oxidation. Generally, the concentration of surfactants in the borohydride alkaline solution should not be increased more than the CMC due to the

241 Chapter 8: Conlusions and Further Work

formation of micelles which can deactivate the electrode and negatively affect the borohydride oxidation decreasing the current density.

The presence of S-228M, and Zonyl FSO in the borohydride solution seem to favour the hydrolysis reaction increasing the hydrogen generation rate at both Au flat plate and

Au/C electrodes which could be useful for hydrogen generation. FC4430 strongly suppresses the borohydride oxidation and its hydrolysis at both planar Au and Au/C electrodes at concentrations of 0.1 wt. % and 0.3 wt. %. Further investigations are required to calculate the CMC of the FC4430 and to analyse its effect at higher and lower concentrations than the CMC.

At the planar Au electrode, the presence of low concentrations of SDS, from 0.00001 wt. % to 0.001 wt. %, seemed to increase both the hydrogen generation and the borohydride oxidation kinetics. At higher concentrations, 0.1 % wt. the limiting current density decreases compared to that in the absence of surfactants and it slightly increased the hydrogen generation rate, mainly at more positive potentials than 0.2 V vs. Hg/HgO.

Although higher currents can be obtained in the presence of low concentrations of SDS, the use of this surfactant it is not recommended for the DBFC at a planar Au anode due to the increase in the rate of hydrogen generation. At the Au/C electrode, 0.1 wt. % SDS decreased the hydrogen generation rate by 4 % with the same limiting current density as in the absence of surfactants. In general, SDS affected the both the hydrolysis and the oxidation of borohydride depending on the concentration.

Triton X-100 decreased the current density and increased the hydrogen generation rate at the planar Au electrode. In contrast, at the Au/C electrode at 0.00001 wt. % - 0.001

242 Chapter 8: Conlusions and Further Work

wt. % concentrations, the hydrogen generation rate decreased from 6 % to 23 % [202], respectively and the polarization curve of borohydride oxidation was not affected. The hydrogen generation rate using 0.1 wt. % Triton X-100 further decreased the hydrogen generation and the oxidation current was much lower than that in the absence of surfactants. Triton X-100 at an optimal concentration, related to the CMC, can be beneficial for the DBFC as has been reported [228, 229]. At concentrations between

0.001 wt. % and 0.01 wt. %, the presence of the surfactant might improve the DBFC performance, by decreasing the gas generation rate, increasing the fuel efficiency and maintaining the power density. At higher concentrations than the CMC, the micelles deactivate the electrode surface minimising both reactions the hydrolysis and the oxidation of borohydride ions.

DFT studies showed that Triton X-100 molecules are most likely to be vertically adsorbed on the Au(111) crystalline face, reducing the formation hydrogen gas molecules from hydrogen atoms from the borohydride hydrolysis or dehydrogenation.

- The results also showed that the free energy of BH4 adsorption is more negative and thus more favourable, when Triton X-100 molecules are already adsorbed.

Experimental and computational studies agree that the presence of Triton X-100 at low coverage (low concentrations) favours the borohydride oxidation and inhibits the borohydride hydrolysis. Further experiments should be carried out in a DBFC with

Triton X-100 in the anolyte solution to confirm the increase in the power density and reduces the gas generation.

243 Chapter 8: Conlusions and Further Work

Mathematical modelling and DFT are powerful tools that can provide a better understanding of the effect of different surfactants on the borohydride oxidation and its hydrolysis, considering the type of surfactant and the different electrode materials.

8.2. Hydrogen generation from sodium borohydride

8.2.1. Hydrogen generation

The glass reactor containing palladium on granulated carbon was the most effective catalyst among all the material tested, with a constant hydrogen generation rate of over

3 -1 3 3 -1 -1 3 600 cm min (8 × 10 cm min gmetal ) during 120 minutes, equivalent to 1359 cm

-1 H2 g NaBH4. This is the highest hydrogen generation rate reported using a Pd based catalyst. Authors such as Peña-Alonso et al. [135] reported higher hydrogen generation rates (90 dm3 min-1 g-1) by doubly extrapolating low values of accumulative hydrogen obtained during a certain period of time (8 cm3 during 20 min). The effect of the quality of water was found to be crucial in the hydrogen generation rate, being appreciable but insufficient when the solution was prepared with river water, probably due to the presence of impurities, and vigorous when the solution was prepared with still drinking water.

By mixing the sodium borohydride powder with the catalyst in the reaction chamber and adding 350 cm3 still drinking water from the fuel chamber, a constant increase in the hydrogen generation rate of approximately 4.7 cm3 min-1 was achieved during at

3 -1 3 3 -1 -1 least 170 min reaching more than 500 cm min (10 × 10 cm min gmetal ). This type of experiment is probably more useful for applications where gradual and constant increase of hydrogen generation rate is needed.

244 Chapter 8: Conlusions and Further Work

SEM images suggested that the palladium catalyst supported on granular carbon loses particles after being used. The EDX analysis shows that 75 % of palladium was lost after three operations probably due to the poor mechanical properties of carbon which disintegrates during operation.

8.3. Suggestions for Further Work

8.3.1. Direct borohydride fuel cell

8.3.1.1. Pd-Ir as anode material for the borohydride oxidation

Promising results were obtained using Pd-Ir as anode material for borohydride oxidation, as very low hydrogen generation rates, while acceptable current densities were obtained (see Chapter 4). However, the following experiments could be performed to further analyze the use of this electrode material for the DBFC:

 Build and test a DBFC using a Pd-Ir anode material, in order to evaluate

the importance and possible inconveniences caused by the generation of

hydrogen gas at the OCV.

 DFT could be used to improve the electrode design. In Chapter 4, Pd2-Ir1

and Pd2-Ir2 alloys were studied using DFT. However, further DFT

analysis could be carried out to improve the electrode design and find a

catalyst composition that minimizes the hydrogen gas generated at the

OCV.

245 Chapter 8: Conlusions and Further Work

8.3.1.2. The presence of surfactants in the electrolyte solution to suppress the

borohydride oxidation

The results shown in Chapter 5 suggested that the presence of surfactants in the electrolyte solution affects the borohydride hydrolysis and its oxidation. The hydrogen generation rate measured during the electrolysis at constant potentials in the presence of concentrations lower than the CMC (0.001 wt. % Triton X-100) decreased the hydrogen generation rate by 23 %, compared to that in the absence of surfactants. Cyclic voltammetry showed that the same surfactant at the same concentration slightly increased the current density obtained from the borohydride oxidation at an Au/C electrode. However, the experiments included in Chapter 5 were performed in half-cells and therefore, the following extension of this work is suggested:

 In order to analyze the effect of the surfactant in the current density and power

density of the DBFC, the building and testing of a DBFC using an electrolyte

that includes Triton X-100 at a concentration between 0.001 wt. % and 0.01

wt. % is recommended.

 The surfactant FC4430 has been barely investigated. It would be useful to

calculate the CMC of FC4430 in water and its variation with the pH of the

solution. This can be easily done by plotting the osmotic pressure or the

conductivity of the solution versus the concentration of surfactant. The CMC

value will be given by the change in the slop; i.e. at lower concentrations that

the CMC, osmotic pressure and the conductivity of the solution will constantly

change versus the concentration of surfactant, whereas at higher concentrations

than the CMC, the osmotic pressure and the conductivity of the solution will not

vary with the concentration of surfactant.

246 Chapter 8: Conlusions and Further Work

 Further studies of the adsorption of surfactants, such as FC4430 and Triton X-

100, on electrode surfaces could be useful to obtain a better understanding of

their effect in the electrode surface and in the borohydride oxidation and

hydrolysis.

8.3.1.3. Gold-coated RVC electrodes for borohydride oxidation

In Chapter 6, it was shown that the current density and the kinetic constants from the borohydride oxidation increased with the deposition time of the gold-coated RVC electrode. It would be possible to extend this work with the investigation of the following:

 Experiments to optimize the deposition time. Once the deposition time has been

optimized, a DBFC could be built with an anode material consisting of the PVD

gold-coated RVC deposited with the optimum time and thickness.

 The experiment previously described could also be modified by adding

surfactants to the electrolyte solution, to decrease the hydrogen generation

during operation and improve the cell performance.

8.3.2. Indirect borohydride fuel cell

Although a considerable hydrogen generation rate has been generated in the glass

3 -1 -1 hydrogen generator using a Pd/C granular catalyst, 6 dm min gcat , the performance of the reactor and the catalyst must be improved prior possible commercialization. The following suggestions for further work are proposed:

247 Chapter 8: Conlusions and Further Work

 The catalyst robustness should be improved via deposition of Pd on a more

robust catalyst, such as Ni.

 The reactor design could be improved by adding a port to feed the solution

separately from the relief valve. The final reactor should be made of stainless

steel, which would stand a higher pressure range. By connecting the reactor to a

H2/O2 fuel cell, efficiency and durability test could be carried out.

 The purity of the hydrogen generated during the hydrolysis reaction should be

quantified. It is necessary to make sure that the silica desiccant is enough to

remove the water steam generated during operation and that the purity of the

hydrogen gas generated is in accordance with the fuel cell specifications.

248 Appendix I

Appendix I. Classification and properties of the surfactants used

Surfactant Type Structure CMC Observations Ref. -5 Zonyl FSO Non-ionic (CF3(CF2)4(EO)10) 9.37 × 10 [230] fluorosurfactant mol dm-3 S-228M Anionic Fluorosurfactant and silicone surfactant bend. Not found High chemical stability. Wide MSDS Fluorosurfactant Composed by hexylene glycol 9.5% and water range of operation in terms of 39% temperature, pH and current density. Triton Non-inonic 0.01 - It is completely stable in liquid [228,

X-100 CH3 CH3 0.03 % formulations containing sodium 229] (1.65 × 10-4 hydroxide (according to the data H3C-C-CH2-C O-(CH2CH2O)N-H - 5 × 10-4 sheet). -3 CH3 CH3 mol dm in aqueous N = approx. 9.5 solution) -3 SDS Anionic NaC12H25SO4 8.2 × 10 [231] surfactant mol dm-3 O O (0.236 S wt.%) in H3C O O Na pure water at 25°C - FC4430 Non-ionic Non-ionic polymeric fluorochemical surfactant Not found Compatible with several resins, [232] Fluorinated coating additives. urethanes, epoxies, polyesters and surfactant Perfluorobutyl (PFB) chemical component is the acrylics. Provide surface tensions building block molecule (4-C perfluorinated superior or comparable to HCs, group) silicone.

249 References

References

1. Zhang, Q., Smith, G., Wu, Y., and Mohring, R., Catalytic hydrolysis of sodium borohydride in an auto-thermal fixed-bed reactor. Int J Hydrogen Energy, 2006. 31(7): p. 961-965. 2. Merino-Jiménez, I., Ponce de León, C., Shah, A.A., and Walsh, F.C., Developments in direct borohydride fuel cells and remaining challenges. Journal of Power Sources, 2012. 219(0): p. 339-357. 3. Rude, L.H., Nielsen, T.K., Ravnsbæk, D.B., Bösenberg, U., Ley, M.B., Richter, B., Arnbjerg, L.M., Dornheim, M., Filinchuk, Y., Besenbacher, F., and Jensen, T.R., Tailoring properties of borohydrides for hydrogen storage: A review. Physica Status Solidi A Appl Res, 2011. 208(8): p. 1754-1773. 4. Ponce de Leon, C., Walsh, F.C., Pletcher, D., Browning, D.J., and Lakeman, J.B., Direct borohydride fuel cells. J Power Sources, 2006. 155(2): p. 172-181. 5. Lewis, B. and von Elbe, G., eds. Combustion, Flames and Explosions of Gases. Third edition ed., ed. I. Academic Press1987: Orlando, Florida. 6. Kuchta, J.M., Investigation of Fire and Explosion Accidents in the Chemical, Mining, and Fuel-Related Industries A Manual. U.S. Bureau of Mines Bulletin, 1985. 680. 7. Demirci, U.B., Direct liquid-feed fuel cells: Thermodynamic and environmental concerns. J. Power Sources, 2007. 169(2): p. 239-246. 8. Lefenfeld, M., Dye, J.L., and Barton, S.C., Sodium Silicide and Alkali Metal- Silica Gel for Convenient Hydrogen Production. Materials Engineering News, Premier Issue Jun (year missing), (Premiere Issue): p. 16-17 (journal discontinued). 9. Liang, Y., Song, K., Shaw, L., Zhu, M., Tait, A., Businelli, N., Yang, J., Soussan, R., Zhou, H., Lu, J., and Mbise, T., The homemaker’s hydrogen generator: A report for IAHE student hydrogen design competition 2010. Int. J. Hydrogen Energy, 2011. 36(21): p. 13880-13887. 10. Retnamma, R., Novais, A.Q., and Rangel, C.M., Kinetics of hydrolysis of sodium borohydride for hydrogen production in fuel cell applications: A review. International Journal of Hydrogen Energy, 2011. 36(16): p. 9772-9790.

250 References

11. Schlapbach, L. and Zuttel, A., Hydrogen-storage materials for mobile applications. Nature, 2001. 414(6861): p. 353-358. 12. Züttel, A., Hydrogen storage methods. Naturwissenschaften 2004. 91(4): p. 157- 172. 13. Lefenfeld, M., Melack, J., and Wallace, P.A., Hydrogen generation systems and methods utilizing sodium silicide and sodium silica gel materials, European patent EP 2414096 A1 2012 Feb 8. 14. Schlesinger, H.I., Brown, H.C., Finholt, A.E., Gilbreath, J.R., Hoekstra, H.R., and Hyde, E.K., Sodium Borohydride, Its Hydrolysis and its Use as a Reducing Agent and in the Generation of Hydrogen. J Am Chem Soc, 1953. 75(1): p. 215- 219. 15. Pinto, A.M.F.R., Ferreira, M.J.F., Fernandes, V.R., and Rangel, C.M., Durability and reutilization capabilities of a Ni–Ru catalyst for the hydrolysis of sodium borohydride in batch reactors. Catal Today, 2011. 170(1): p. 40-49. 16. Kojima, Y., Suzuki, K.-i., Fukumoto, K., Sasaki, M., Yamamoto, T., Kawai, Y., and Hayashi, H., Hydrogen generation using sodium borohydride solution and metal catalyst coated on metal oxide. International Journal of Hydrogen Energy, 2002. 27(10): p. 1029-1034. 17. Feng, R.X., Dong, H., Wang, Y.D., Ai, X.P., Cao, Y.L., and Yang, H.X., A

simple and high efficient direct borohydride fuel cell with MnO2-catalyzed cathode. Electrochem. Comm., 2005. 7(4): p. 449-452. 18. Miley, G.H., Luo, N., Mather, J., Burton, R., Hawkins, G., Gu, L., Byrd, E., Gimlin, R., Shrestha, P.J., Benavides, G., Laystrom, J., and Carroll, D., Direct

NaBH4/H2O2 fuel cells. J Power Sources, 2007. 165(2): p. 509-516. 19. Martins, J.I., Nunes, M.C., Koch, R., Martins, L., and Bazzaoui, M., Electrochemical oxidation of borohydride on platinum electrodes: The influence of thiourea in direct fuel cells. Electrochim Acta, 2007. 52(23): p. 6443-6449. 20. Finkelstein, D. ., Mota, N.D., Cohen, . ., and bru a, H.c.D., Rotating Disk − − Electrode (RDE) Investigation of BH4 and BH3OH Electro-oxidation at Pt − and Au: Implications for BH4 Fuel Cells. J. Phys. Chem. C, 2009. 113(45): p. 19700-19712.

251 References

21. Demirci, U.B., Direct borohydride fuel cell: Main issues met by the membrane– electrodes-assembly and potential solutions. J. Power Sources, 2007. 172(2): p. 676-687. 22. Kojima, Y. and Haga, T., Recycling process of sodium metaborate to sodium borohydride. Int J Hydrogen Energy, 2003. 28(9): p. 989-993. 23. Ponce de León, C., Walsh, F.C., Rose, A., Lakeman, J.B., Browning, D.J., and Reeve, R.W., A direct borohydride--Acid peroxide fuel cell. Journal of Power Sources, 2007. 164(2): p. 441-448. 24. Wee, J.-H., Which type of fuel cell is more competitive for portable application: Direct methanol fuel cells or direct borohydride fuel cells? Journal of Power Sources, 2006. 161(1): p. 1-10. 25. Grinberg, V., Mayorova, N., and Pasynskii, A., A cluster Pt-Sn-catalyst for the ethanol direct oxidation. Russian Journal of Electrochemistry, 2009. 45(12): p. 1321-1326. 26. Lamy, C., Coutanceau, C., and Leger, J.M., The Direct Ethanol Fuel Cell: a Challenge to Convert Bioethanol Cleanly into Electric Energy. Catalysis for Sustainable Energy Production, 2009. 27. Elder, J.P. and Hickling, A., Anodic behaviour of the borohydride ion. Trans. Faraday Soc., 1962. 58: p. 1852-1864. 28. Indig, M.E. and Snyder, R.N., Sodium borohydride, an interesting anodic fuel. J. Electroanal. Chem., 1962. 109. 29. Santos, D.M.F. and Sequeira, C.A.C., Sodium borohydride as a fuel for the future. Renew. Sust. Energy Rev., 2011. 15(8): p. 3980-4001. 30. Gyenge, E., Atwan, M., and Northwood, D., Electrocatalysis of Borohydride Oxidation on Colloidal Pt and Pt-Alloys (Pt-Ir, Pt-Ni, and Pt-Au) and Application for Direct Borohydride Fuel Cell Anodes. J. Electrochem. Soc., 2006. 153(1): p. A150-A158. 31. Liu, B.H., Li, Z.P., and Suda, S., Electrocatalysts for the anodic oxidation of borohydrides. Electrochim. Acta, 2004. 49(19): p. 3097-3105. 32. Gyenge, E., Electrooxidation of borohydride on platinum and gold electrodes: implications for direct borohydride fuel cells. Electrochim. Acta, 2004. 49(6): p. 965-978.

252 References

33. Mirkin, M.V., Yang, H., and Bard, A.J., Borohydride Oxidation at a Gold Electrode. J. Electrochem. Soc., 1992. 139(8): p. 2212-2217. 34. Lee, S.M., Kim, J.H., Lee, H.H., Lee, P.S., and Lee, J.Y., The Characterization of an Alkaline Fuel Cell That Uses Hydrogen Storage Alloys. J. Electrochem. Soc., 2002. 149(5): p. A603-A606. 35. Morris, J.H., Gysling, H.J., and Reed, D., Electrochemistry of boron compounds. Chem. Rev., 1985. 85(1): p. 51-76. 36. Gu, L., Luo, N., and Miley, G.H., Cathode electrocatalyst selection and deposition for a direct borohydride/hydrogen peroxide fuel cell. J. Power Sources, 2007. 173(1): p. 77-85. 37. Escaño, M.C.S., Gyenge, E., Arevalo, R.L., and Kasai, H., Reactivity Descriptors for Borohydride Interaction with Metal Surfaces. The Journal of Physical Chemistry C, 2011. 115(40): p. 19883-19889. 38. Duteanu, N., Vlachogiannopoulos, G., Shivhare, M., Yu, E., and Scott, K., A parametric study of a platinum ruthenium anode in a direct borohydride fuel cell. J. Applied Electrochem., 2007. 37(9): p. 1085-1091. 39. Amendola, S.C., Onnerud, P., Kelly, M.T., Petillo, P.J., Sharp-Goldman, S.L., and Binder, M., A novel high power density borohydride-air cell. J. Power Sources, 1999. 84(1): p. 130-133. 40. Liu, B.H. and Li, Z.P., Current status and progress of direct borohydride fuel cell technology development. J. Power Sources, 2009. 187(2): p. 291-297. 41. Chatenet, M., Micoud, F., Roche, I., Chainet, E., and Vondrák, J., Kinetics of sodium borohydride direct oxidation and oxygen reduction in sodium hydroxide

electrolyte: Part II. O2 reduction. Electrochim. Acta, 2006. 51(25): p. 5452- 5458. 42. Atwan, M.H., Northwood, D.O., and Gyenge, E.L., Evaluation of colloidal Os and Os-Alloys (Os–Sn, Os–Mo and Os–V) for electrocatalysis of methanol and borohydride oxidation. Int. J. Hydrogen Energy, 2005. 30(12): p. 1323-1331. 43. Bessette, R.R., Medeiros, M.G., Patrissi, C.J., Deschenes, C.M., and LaFratta, C.N., Development and characterization of a novel carbon fiber based cathode for semi-fuel cell applications. J. Power Sources, 2001. 96(1): p. 240-244.

253 References

44. Bessette, R.R., Cichon, J.M., Dischert, D.W., and Dow, E.G., A study of cathode catalysis for the aluminium/hydrogen peroxide semi-fuel cell. J. Power Sources, 1999. 80(1-2): p. 248-253.

45. Luo, N., Miley, G.H., Kim, K.-J., Burton, R., and Huang, X., NaBH4/H2O2 fuel cells for air independent power systems. J. Power Sources, 2008. 185(2): p. 685- 690. 46. Li, Z.P., Liu, B.H., Arai, K., and Suda, S., A Fuel Cell Development for Using Borohydrides as the Fuel. J. Electrochem. Soc., 2003. 150(7): p. A868-A872. 47. Choudhury, N.A., Raman, R.K., Sampath, S., and Shukla, A.K., An alkaline direct borohydride fuel cell with hydrogen peroxide as oxidant. J. Power Sources, 2005. 143(1-2): p. 1-8. 48. Li, Z.P., Liu, B.H., Arai, K., Asaba, K., and Suda, S., Evaluation of alkaline borohydride solutions as the fuel for fuel cell. J. Power Sources, 2004. 126(1-2): p. 28-33. 49. Cheng, H., Scott, K., and Lovell, K., Material Aspects of the Design and Operation of Direct Borohydride Fuel Cells. Fuel Cells, 2006. 6(5): p. 367-375. 50. Cheng, H. and Scott, K., Investigation of Ti mesh-supported anodes for direct borohydride fuel cells. J. Applied Electrochem., 2006. 36(12): p. 1361-1366. 51. Molina Concha, B. and Chatenet, M., Direct oxidation of sodium borohydride on Pt, Ag and alloyed Pt–Ag electrodes in basic media: Part II. Carbon- supported nanoparticles. Electrochim. Acta, 2009. 54(26): p. 6130-6139. 52. Rostamikia, G. and Janik, M.J., Direct borohydride oxidation: mechanism determination and design of alloy catalysts guided by density functional theory. Energy Enviro. Sci., 2010. 3(9): p. 1262-1274. 53. Tegou, A., Armyanov, S., Valova, E., Steenhaut, O., Hubin, A., Kokkinidis, G., and Sotiropoulos, S., Mixed platinum–gold electrocatalysts for borohydride oxidation prepared by the galvanic replacement of nickel deposits. Journal of Electroanalytical Chemistry, 2009. 634(2): p. 104-110. 54. Concha, B.M., Chatenet, M., Ticianelli, E.A., Lima, F.H., and Lima, R.D., In Situ Infrared Study of the Mechanism of Borohydride Oxidation Reaction. ECS Meeting Abstracts, 2009. 902(6): p. 368.

254 References

55. Santos, D.M.F. and Sequeira, C.A.C., Cyclic voltammetry investigation of borohydride oxidation at a gold electrode. Electrochim. Acta, 2010. 55(22): p. 6775-6781. 56. Chatenet, M., Lima, F.H.B., and Ticianelli, E.A., Gold is not a Faradaic- Efficient Borohydride Oxidation Electrocatalyst: An Online Electrochemical Mass Spectrometry Study. J. Electrochem. Soc., 2010. 157(5): p. B697-B704. 57. He, P., Wang, X., Fu, P., Wang, H., and Yi, L., The studies of performance of the Au electrode modified by Zn as the anode electrocatalyst of direct borohydride fuel cell. International Journal of Hydrogen Energy, 2011. 36(15): p. 8857-8863. 58. Yi, L., Song, Y., Liu, X., Wang, X., Zou, G., He, P., and Yi, W., High activity of Au–Cu/C electrocatalyst as anodic catalyst for direct borohydride-hydrogen peroxide fuel cell. Int. J. Hydrogen Energy, 2011. 36(24): p. 15775-15782. 59. Pei, F., Wang, Y., Wang, X., He, P., Chen, Q., Wang, X., Wang, H., Yi, L., and Guo, J., Performance of supported Au–Co alloy as the anode catalyst of direct borohydride-hydrogen peroxide fuel cell. Int. J. Hydrogen Energy, 2010. 35(15): p. 8136-8142. 60. Atwan, M.H., Northwood, D.O., and Gyenge, E.L., Evaluation of colloidal Ag and Ag-alloys as anode electrocatalysts for direct borohydride fuel cells. Int. J. Hydrogen Energy, 2007. 32(15): p. 3116-3125. 61. Sanli, E., Celikkan, H., Uysal, B.Z., and Aksu, M.L., Impedance Analysis and Electrochemical Measurements of A Direct Borohydride Fuel Cell Constructed with Ag Anode. ECS Trans., 2007. 5(1): p. 137-145. 62. Sanlı, E., Çelikkan, H., Zühtü Uysal, B., and ksu, M.L., Anodic behavior of Ag metal electrode in direct borohydride fuel cells. Int. J. Hydrogen Energy, 2006. 31(13): p. 1920-1924. 63. Liu, B.H., Li, Z.P., and Suda, S., Anodic Oxidation of Alkali Borohydrides Catalyzed by Nickel. J. Electrochem. Soc., 2003. 150(3): p. A398-A402. 64. Ma, J., Sahai, Y., and Buchheit, R.G., Direct borohydride fuel cell using Ni- based composite anodes. J. Power Sources, 2010. 195(15): p. 4709-4713. 65. Santos, D.M. and Sequeira, C.A., Zinc Negative Electrode for Direct Borohydride Fuel Cells. ECS Trans., 2009. 16(50): p. 123-137.

255 References

66. McCafferty, E., Introduction to corrosion sciences2010, New York, USA: Springer. 67. Celik, C., Boyaci San, F.G., and Sarac, H.I., Effects of operation conditions on direct borohydride fuel cell performance. J. Power Sources, 2008. 185(1): p. 197-201. 68. Yi, L., Song, Y., Wang, X., Yi, L., Hu, J., Su, G., Yi, W., and Yan, H., Carbon supported palladium hollow nanospheres as anode catalysts for direct borohydride-hydrogen peroxide fuel cells. Journal of Power Sources, 2012. 205(0): p. 63-70. 69. Atwan, M.H., Gyenge, E., and Northwood, D.O., Evaluation of Colloidal Pd and Pd-alloys as anode electrocatalysts for direct borohydride fuel cells applications. Journal of New Materials for Electrochemical Systems, 2010. 13(1): p. 21-27. 70. Kiran, V., Ravikumar, T., Kalyanasundaram, N.T., Krishnakumar, S., Shukla, A.K., and Sampath, S., Electro-Oxidation of Borohydride on Rhodium, Iridium, and Rhodium–Iridium Bimetallic Nanoparticles with Implications to Direct Borohydride Fuel Cells J. Electrochem. Soc., 2010. 157(8): p. B1201-B1208. 71. Zhi-fang, J., Dong-hong, D., and Yan-ping, S., The electrochemical behaviour - of alkaline BH4 on copper anode. Battery Biomonthly, 2008. 133. 72. Hampton, M.D., Shur, D.V., Zaginaichenko, S.Y., and Trefilov, V.I., Hydrogen Materials Science and Chemistry of Metal Hydrides2002, Dodrecht, The Netherlands: Kluwer Academic Publishers. 73. Kiehne, H.A., ed. Battery Technology Handbook, (second ed.), Expert Verlag GmbH, Renningen-Malsheim, Germany 2003.

74. Wang, L., Ma, C., Sun, Y., and Suda, S., AB5-type hydrogen storage alloy used as anodic materials in borohydride fuel cell. J. Alloys Compounds, 2005. 391(1- 2): p. 318-322.

75. Wang, L., Ma, C.-a., and Mao, X., LmNi4.78Mn0.22 alloy modified with Si used as anodic materials in borohydride fuel cells. J. Alloys Compounds, 2005. 397(1- 2): p. 313-316. 76. Selvarani, G., Prashant, S.K., Sahu, A.K., Sridhar, P., Pitchumani, S., and Shukla, A.K., A direct borohydride fuel cell employing Prussian Blue as

256 References

mediated electron-transfer hydrogen peroxide reduction catalyst. J. Power Sources, 2008. 178(1): p. 86-91. 77. Ma, J., Liu, Y., Zhang, P., and Wang, J., A simple direct borohydride fuel cell with a cobalt phthalocyanine catalyzed cathode. Electrochem. Comm., 2008. 10(1): p. 100-102. 78. Raman, R.K. and Shukla, A.K., A Direct Borohydride/Hydrogen Peroxide Fuel Cell with Reduced Alkali Crossover. Fuel Cells, 2007. 7(3): p. 225-231. 79. Raman, R.K. and Shukla, A.K., Electro-reduction of hydrogen peroxide on iron tetramethoxy phenyl porphyrin and lead sulfate electrodes with application in direct borohydride fuel cells. J. Applied Electrochem., 2005. 35(11): p. 1157- 1161. 80. Raman, R.K., Prashant, S.K., and Shukla, A.K., A 28-W portable direct borohydride–hydrogen peroxide fuel-cell stack. J. Power Sources, 2006. 162(2): p. 1073-1076. 81. Liu, B.H. and Suda, S., Hydrogen storage alloys as the anode materials of the direct borohydride fuel cell. J. Alloys Compounds, 2008. 454(1-2): p. 280-285. 82. Li, Z.P., Liu, B.H., Arai, K., and Suda, S., Development of the direct borohydride fuel cell. J. Alloys Compounds, 2005. 404-406(0): p. 648-652. 83. Özkar, S. and Zahmakıran, M., Hydrogen generation from hydrolysis of sodium borohydride using Ru(0) nanoclusters as catalyst. J. Alloys Compounds, 2005. 404–406(0): p. 728-731. 84. Wu, C., Wu, F., Bai, Y., Yi, B., and Zhang, H., Cobalt boride catalysts for

hydrogen generation from alkaline NaBH4 solution. Mater. Lett., 2005. 59(14- 15): p. 1748-1751. 85. Strasser, K., The design of alkaline fuel cells. J. Power Sources, 1990. 29(1-2): p. 149-166. 86. Ponce de Leon, C., Walsh, F.C., Patrissi, C.J., Medeiros, M.G., Bessette, R.R., Reeve, R.W., Lakeman, J.B., Rose, A., and Browning, D., A direct borohydride- peroxide fuel cell using a Pd/Ir alloy coated microfibrous carbon cathode. Electrochem. Comm., 2008. 10(10): p. 1610-1613. 87. Cheng, H. and Scott, K., Investigation of non-platinum cathode catalysts for direct borohydride fuel cells. J. Electroanal. Chem., 2006. 596(2): p. 117-123.

257 References

88. Cheng, H. and Scott, K., Influence of operation conditions on direct borohydride fuel cell performance. J. Power Sources, 2006. 160(1): p. 407-412. 89. Ponce de León, C., Walsh, F.C., Rose, A., Lakeman, J.B., Browning, D.J., and Reeve, R.W., A direct borohydride—Acid peroxide fuel cell. J. Power Sources, 2007. 164(2): p. 441-448. 90. Verma, A., Jha, A.K., and Basu, S., Manganese dioxide as a cathode catalyst for a direct alcohol or sodium borohydride fuel cell with a flowing alkaline electrolyte. J. Power Sources, 2005. 141(1): p. 30-34. 91. Cheng, H., Scott, K., Lovell, K.V., Horsfall, J.A., and Waring, S.C., Evaluation of new ion exchange membranes for direct borohydride fuel cells. J. Membrane Sci., 2007. 288(1-2): p. 168-174. 92. Ma, J., Choudhury, N.A., Sahai, Y., and Buchheit, R.G., Performance Study of Direct Borohydride Fuel Cells Employing Polyvinyl Alcohol Hydrogel Membrane and Nickel-Based Anode. Fuel Cells, 2011. 11(5): p. 603-610. 93. Pletcher, D., ed. A First Course in Electrode Processes. second ed. 2009, RSC. 94. Walsh, F., A First Course in Electrochemical Engineering, ed. T.E. Consultancy1993, Cambridge, UK: RSC Publishing. 95. Wang, K., Lu, J., and Zhuang, L., A mechanistic study of borohydride anodic oxidation. Catal. Today, 2011. 170(1): p. 99-109. 96. Li, Z.P., Liu, B.H., Zhu, J.K., and Suda, S., Depression of hydrogen evolution during operation of a direct borohydride fuel cell. J. Power Sources, 2006. 163(1): p. 555-559. 97. Demirci, Ü.B., Comments on the paper "Electrooxidation of borohydride on platinum and gold electrodes: Implications for direct borohydride fuel cell" by E. Gyenge, Electrochim. Acta 49 (2004) 965: Thiourea, a poison for the anode metallic electrocatalyst of the direct borohydride fuel cell? Electrochimica Acta, 2007. 52(15): p. 5119-5121. 98. Jamard, R., Latour, A., Salomon, J., Capron, P., and Martinent-Beaumont, A., Study of fuel efficiency in a direct borohydride fuel cell. J. Power Sources, 2008. 176(1): p. 287-292. 99. Celik, C., Boyaci San, F.G., and Sarac, H.I., Improving the direct borohydride fuel cell performance with thiourea as the additive in the sodium borohydride solution. Int. J. Hydrogen Energy, 2010. 35(16): p. 8678-8682.

258 References

100. Dang, X., Wei, Y., and Hu, S., Effects of Surfactants on the Electroreduction of Dioxygen at an Acetylene Black Electrode. Anal Sci, 2004. 20: p. 307-310. 101. Vittal, R., Gomathi, H., and Kim, K.-J., Beneficial role of surfactants in electrochemistry and in the modification of electrodes. Adv. Colloid Interface Sci., 2006. 119(1): p. 55-68. 102. He, Q., Yuan, S., Chen, C., and Hu, S., Electrochemical properties of estradiol at glassy carbon electrode modified with nano-Al2O3 film. Mater. Sci. Eng. C Mater. Biol. Appl., 2003. 23(5): p. 621-625. 103. Ferreira, M.J.F., Fernandes, V.R., Gales, L., Rangel, C.M., and Pinto, A.M.F.R., Effects of the addition of an organic polymer on the hydrolysis of sodium tetrahydroborate in batch reactors. Int J Hydrogen Energy, 2010. 35(20): p. 11456-11469. 104. Umit B, D., Direct liquid-feed fuel cells: Thermodynamic and environmental concerns. J. Power Sources, 2007. 169(2): p. 239-246. 105. Zhang, Q., Smith, G.M., and Wu, Y., Catalytic hydrolysis of sodium borohydride in an integrated reactor for hydrogen generation. Int J Hydrogen Energy, 2007. 32(18): p. 4731-4735. 106. Celik, C., San, F.G.B., and Sarac, H.I., Influences of sodium borohydride concentration on direct borohydride fuel cell performance. J. Power Sources, 2010. 195(9): p. 2599-2603. 107. Kim, S.J., Lee, J., Kong, K.Y., Ryul Jung, C., Min, I.-G., Lee, S.-Y., Kim, H.-J., Nam, S.W., and Lim, T.-H., Hydrogen generation system using sodium borohydride for operation of a 400W-scale polymer electrolyte fuel cell stack. J. Power Sources, 2007. 170(2): p. 412-418. 108. Kim, C., Kim, K.-J., and Ha, M.Y., Performance enhancement of a direct borohydride fuel cell in practical running conditions. J. Power Sources, 2008. 180(1): p. 154-161. 109. Ponce de León, C., Kulak, A., Williams, S., Merino-Jiménez, I., and Walsh, F.C., Improvements in direct borohydride fuel cells using three-dimensional electrodes. Catal. Today, 2011. 170(1): p. 148-154. 110. Ma, J., Choudhury, N.A., and Sahai, Y., A comprehensive review of direct borohydride fuel cells. Renew. Sust. Energy Rev., 2010. 14(1): p. 183-199.

259 References

111. Yang, H.F., Rao, G.H., Liu, G.Y., Ouyang, Z.W., Liu, W.F., Feng, X.M., Chu, W.G., and Liang, J.K., Crystal Structures of Compounds in the Pseudobinary System Gd5Ge4—La5Ge4. ChemInform, 2004. 35(1): p. no-no. 112. Yang, H. and Zhao, T.S., Effect of anode flow field design on the performance of liquid feed direct methanol fuel cells. Electrochim. Acta, 2005. 50(16-17): p. 3243-3252. 113. Park, K.T., Jung, U.H., Jeong, S.U., and Kim, S.H., Influence of anode diffusion layer properties on performance of direct borohydride fuel cell. J. Power Sources, 2006. 162(1): p. 192-197. 114. Qian, W., Wilkinson, D.P., Shen, J., Wang, H., and Zhang, J., Architecture for portable direct liquid fuel cells. J. Power Sources, 2006. 154(1): p. 202-213. 115. Kim, C., Kim, K.-J., and Ha, M.Y., Investigation of the characteristics of a stacked direct borohydride fuel cell for portable applications. J. Power Sources, 2008. 180(1): p. 114-121. 116. Verma, A. and Basu, S., Experimental evaluation and mathematical modeling of a direct alkaline fuel cell. J. Power Sources, 2007. 168(1): p. 200-210. 117. Sanli, A.E., Aksu, M.L., and Uysal, B.Z., Advanced mathematical model for the passive direct borohydride/peroxide fuel cell. Int. J. Hydrogen Energy, 2011. 36(14): p. 8542-8549. 118. Shah, A.A., Singh, R., Ponce de León, C., Wills, R.G., and Walsh, F.C., Mathematical modelling of direct borohydride fuel cells. Journal of Power Sources, 2013. 221(0): p. 157-171. 119. Rostamikia, G. and Janik, M.J., Borohydride Oxidation over Au(111): A First- Principles Mechanistic Study Relevant to Direct Borohydride Fuel Cells. J. Electrochem. Soc., 2009. 156(1): p. B86-B92. 120. Rostamikia, G. and Janik, M.J., First principles mechanistic study of borohydride oxidation over the Pt(111) surface. Electrochim. Acta, 2010. 55(3): p. 1175-1183. 121. Rostamikia, G., Mendoza, A.J., Hickner, M.A., and Janik, M.J., First-principles based microkinetic modeling of borohydride oxidation on a Au(111) electrode. J. Power Sources, 2011. 196(22): p. 9228-9237.

260 References

122. Rostamikia, G., Patel, R.J., Merino-Jimenez, I., Hickner, M., and Janik, M.J., Electrocatalyst design for direct borohydride oxidation guided by first principles. Ready for submission, 2013. 123. Wu, Y., Kelly, M.T., and Ortega, J., Review of chemical processes for the synthesis of sodium borohydride. Millenium Cell Inc., 2004. 124. Sanli, .E., Kayacan, İ., Uysal, B.Z., and ksu, M. ., Recovery of borohydride from metaborate solution using a silver catalyst for application of direct rechargable borohydride/peroxide fuel cells. J. Power Sources, 2010. 195(9): p. 2604-2607. 125. Gyenge, E.L. and Oloman, C.W., Electrosynthesis attempts of tetrahydridoborates. J. Applied Electrochem., 1998. 28(10): p. 1147-1151. 126. Cooper, H.B.H., Electrolytic Process for the Production of alkali Borohydrides, 1973. 127. Sharifian, H. and Dutcher, J.S., Production of Quaternary Ammonium and Quaternary Phosphonium Borohydrides, 1990. 128. Hale, C.H. and Sharifian, H., Production of Metal Borohydride and Organic Onium Borohydrides, 1990. 129. Sun, Y. and Liang, Z., Electrochemical Process for Preparing Borohydride, 2003. 130. Santos, D.M.F. and Sequeira, C.A.C., On the electrosynthesis of sodium borohydride. International Journal of Hydrogen Energy, 2010. 35(18): p. 9851- 9861. 131. Broja G and W., S., Process for the production of alkali metal borohydrides, D. Patent, Editor 1959. 132. Martelli, P., Caputo, R., Remhof, A., Mauron, P., Borgschulte, A., and Zu¨ ttel,

A., Stability and Decomposition of NaBH4. The Journal of Physical Chemistry C, 2010. 15(114): p. 7173-7177. 133. Li, Z.P., Morigazaki, N., Liu, B.H., and Suda, S., Preparation of sodium

borohydride by the reaction of MgH2 with dehydrated borax through ball milling at room temperature. Journal of Alloys and Compounds, 2003. 349(1– 2): p. 232-236.

261 References

134. Marrero-Alfonso, E.Y., Beaird, A.M., Davis, T.A., and Matthews, M.A., Hydrogen Generation from Chemical Hydrides. Ind Eng Chem Res, 2009. 48(8): p. 3703-3712. 135. Peña-Alonso, R., Sicurelli, A., Callone, E., Carturan, G., and Raj, R., A

picoscale catalyst for hydrogen generation from NaBH4 for fuel cells. J Power Sources, 2007. 165(1): p. 315-323. 136. Guella, G., Zanchetta, C., Patton, B., and Miotello, A., New Insights on the Mechanism of Palladium-Catalyzed Hydrolysis of Sodium Borohydride from 11B NMR Measurements. J Phys Chem B, 2006. 110(34): p. 17024-17033. 137. Liu, B.H. and Li, Z.P., A review: Hydrogen generation from borohydride hydrolysis reaction. J. Power Sources, 2009. 187(2): p. 527-534. 138. Liu, B.H., Li, Z.P., and Suda, S., Solid sodium borohydride as a hydrogen source for fuel cells. J Alloys Compd, 2009. 468(1-2): p. 493-498. 139. Muir, S.S. and Yao, X., Progress in sodium borohydride as a hydrogen storage material: Development of hydrolysis catalysts and reaction systems. Int J Hydrogen Energy, 2011. 36(10): p. 5983-5997. 140. Wu, C., Zhang, H., and Yi, B., Hydrogen generation from catalytic hydrolysis of sodium borohydride for proton exchange membrane fuel cells. Catal Today, 2004. 93-95(0): p. 477-483. 141. Bai, Y., Wu, C., Wu, F., and Yi, B., Carbon-supported platinum catalysts for

on-site hydrogen generation from NaBH4 solution. Mater Lett, 2006. 60(17-18): p. 2236-2239. 142. Liu, Z., Guo, B., Chan, S.H., Tang, E.H., and Hong, L., Pt and Ru dispersed on

LiCoO2 for hydrogen generation from sodium borohydride solutions. J. Power Sources, 2008. 176(1): p. 306-311. 143. Park, J.-H., Shakkthivel, P., Kim, H.-J., Han, M.-K., Jang, J.-H., Kim, Y.-R., Kim, H.-S., and Shul, Y.-G., Investigation of metal alloy catalyst for hydrogen release from sodium borohydride for polymer electrolyte membrane fuel cell application. Int. J. Hydrogen Energy, 2008. 33(7): p. 1845-1852. 144. Liang, Y., Dai, H.-B., Ma, L.-P., Wang, P., and Cheng, H.-M., Hydrogen generation from sodium borohydride solution using a ruthenium supported on graphite catalyst. Int. J. Hydrogen Energy, 2010. 35(7): p. 3023-3028.

262 References

145. Oezkar, S. and Zahmakiran, M., Hydrogen Generation from Hydrolysis of Sodium Borohydride Using Ru(0) Nanoclusters as Catalyst. ChemInform, 2006. 37(11): p. no-no. 146. Zahmakıran, M. and Özkar, S., Water dispersible acetate stabilized ruthenium(0) nanoclusters as catalyst for hydrogen generation from the hydrolysis of sodium borohyride. J Mol Catal A Chem, 2006. 258(1-2): p. 95- 103. 147. Liu, B.H. and Li, Q., A highly active Co-B catalyst for hydrogen generation from sodium borohydride hydrolysis. Int. J. Hydrogen Energy, 2008. 33(24): p. 7385-7391. 148. Liang, Y., Wang, P., and Dai, H.-B., Hydrogen bubbles dynamic template preparation of a porous Fe–Co–B/Ni foam catalyst for hydrogen generation from hydrolysis of alkaline sodium borohydride solution. J. Alloys Compd, 2010. 491(1-2): p. 359-365. 149. Dai, H.B., Liang, Y., Wang, P., Yao, X.D., Rufford, T., Lu, M., and Cheng, H.M., High-performance cobalt–tungsten–boron catalyst supported on Ni foam for hydrogen generation from alkaline sodium borohydride solution. Int. J. Hydrogen Energy, 2008. 33(16): p. 4405-4412. 150. Amendola, S.C., Sharp-Goldman, S.L., Janjua, M.S., Spencer, N.C., Kelly, M.T., Petillo, P.J., and Binder, M., A safe, portable, hydrogen gas generator using aqueous borohydride solution and Ru catalyst. Int J Hydrogen Energy, 2000. 25(10): p. 969-975. 151. Kojima, Y., Suzuki, K.-i., Fukumoto, K., Kawai, Y., Kimbara, M., Nakanishi, H., and Matsumoto, S., Development of 10 kW-scale hydrogen generator using chemical hydride. J. Power Sources, 2004. 125(1): p. 22-26. 152. Murugesan, S. and Subramanian, V., Effects of acid accelerators on hydrogen generation from solid sodium borohydride using small scale devices. J. Power Sources, 2009. 187(1): p. 216-223. 153. Prosini, P.P. and Gislon, P., A hydrogen refill for cellular phone. J. Power Sources, 2006. 161(1): p. 290-293. 154. Akdim, O., Demirci, U.B., and Miele, P., Acetic acid, a relatively green single- use catalyst for hydrogen generation from sodium borohydride. Int. J. Hydrogen Energy, 2009. 34(17): p. 7231-7238.

263 References

155. Amedola, S., Brinder, M., Sharp-Goldman, S., Kelly, M., and Petillo, P., Process for making hydrogen generation catalyst, 2004. 156. Zhang, J., Zheng, Y., Gore, J.P., and Fisher, T.S., 1 kWe sodium borohydride hydrogen generation system: Part I: Experimental study. J. Power Sources, 2007. 165(2): p. 844-853. 157. Zhang, Q., Smith, G., and Hong, Z., Catalytic reactor for hydrogen generation systems, in US patent application2005. 158. Kim, K., Kim, T., Lee, K., and Kwon, S., Fuel cell system with sodium borohydride as hydrogen source for unmanned aerial vehicles. J. Power Sources, 2011. 196(21): p. 9069-9075. 159. Xia, Z.T. and Chan, S.H., Feasibility study of hydrogen generation from sodium borohydride solution for micro fuel cell applications. J. Power Sources, 2005. 152(0): p. 46-49. 160. Marrero-Alfonso, E.Y., Gray, J.R., Davis, T.A., and Matthews, M.A., Minimizing water utilization in hydrolysis of sodium borohydride: The role of sodium metaborate hydrates. Int. J. Hydrogen Energy, 2007. 32(18): p. 4723- 4730.

161. Shang, Y. and Chen, R., Hydrogen Storage via the Hydrolysis of NaBH4 Basic

Solution: Optimization of NaBH4 Concentration. Energy Fuels, 2006. 20(5): p. 2142-2148. 162. Kojima, Y., Suzuki, K.-i., Fukumoto, K., Sasaki, M., Yamamoto, T., Kawai, Y., and Hayashi, H., Hydrogen generation using sodium borohydride solution and metal catalyst coated on metal oxide. Int. J. Hydrogen Energy, 2002. 27(10): p. 1029-1034. 163. Alibeyli, R., Arslan, S., and Ozdemir, E., Single stage production and hydrolysis of sodium borohydride. Int. J. Hydrogen Energy, 2011. 36(Copyright (C) 2012 American Chemical Society (ACS). All Rights Reserved.): p. 11451-11456. 164. Demirci, U.B., Comments on the paper “Electrooxidation of borohydride on platinum and gold electrodes: Implications for direct borohydride fuel cell” by E. Gyenge, Electrochim. Acta 49 (2004) 965: Thiourea, a poison for the anode metallic electrocatalyst of the direct borohydride fuel cell? Electrochim. Acta, 2007. 52(15): p. 5119-5121.

264 References

165. Patrissi, C.J., Bessette, R.R., Kim, Y.K., and Schumacher, C.R., Fabrication and

Rate Performance of a Microfiber Cathode in a Mg--H2O2 Flowing Electrolyte Semi-Fuel Cell. J Electrochem Soc, 2008. 155(6): p. B558-B562. 166. Litz, L.M., Lakewood., and Rothfleisch, J.E., Hydrogen generator, United States patent US 3459510, 1969 Aug 5: United States. 167. Halsey, G., Physical Adsorption on Non-Uniform Surfaces J Chem Phys, 1948. 16(10): p. 931-937. 168. Bolzan, A.E., Phenomenological aspects related to the electrochemical behaviour of smooth palladium electrodes in alkaline solutions. J. Electroanal. Chem., 1995. 380: p. 127-138. 169. Denuault, G., Milhano, C., and Pletcher, D., Mesoporous palladium-the surface electrochemistry of palladium in aqueous sodium hydroxide and the cathodic reduction of nitrite. Phys. Chem. Chem. Phys., 2005. 7: p. 3545-3551. 170. Liang, Z.X., Zhao, T.S., Xu, J.B., and Zhu, L.D., Mechanism study of the ethanol oxidation reaction on palladium in alkaline media. Electrochimica Acta, 2009. 54(8): p. 2203-2208. 171. Anderson, A.B. and Kang, D.B., Quantum Chemical Approach to Redox Reactions Including Potential Dependence: Application to a Model for Hydrogen Evolution from Diamond. The Journal of Physical Chemistry A, 1998. 102(29): p. 5993-5996. 172. Nørskov, J.K., Rossmeisl, J., Logadottir, A., Lindqvist, L., Kitchin, J.R., Bligaard, T., and Jónsson, H., Origin of the Overpotential for Oxygen Reduction at a Fuel-Cell Cathode. The Journal of Physical Chemistry B, 2004. 108(46): p. 17886-17892. 173. Kresse, G. and Furthmüller, J., Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Computational Materials Science, 1996. 6(1): p. 15-50. 174. Kresse, G. and Furthmüller, J., Efficient iterative schemes for ab initio total- energy calculations using a plane-wave basis set. Phys. Rev. B, 1996. 54: p. 11169. 175. Kresse, G. and Hafner, J., Ab initio molecular dynamics for liquid metals. Phys. Rev. B, 1993. 47: p. 558. 176. http://www.vasp.at/. 2009.

265 References

177. Nie, X., Esopi, M.R., Janik, M.J., and Asthagiri, A., Selectivity of CO2 Reduction on Copper Electrodes: The Role of the Kinetics of Elementary Steps. Angew. Chem. Int., 2013. 52: p. 2459-2462. 178. Wenger, D., Polifke, W., Schmidt-Ihn, E., Abdel-Baset, T., and Maus, S., Comments on solid state hydrogen storage systems design for fuel cell vehicles. International Journal of Hydrogen Energy, 2009. 34(15): p. 6265-6270. 179. Finocchi, F., ed. Density Functional Theory for Beginners, Basic Principles and Practical Approaches. 2011, Institut des NanoSciences de Paris (INSP), CNRS and University Pierre et Marie Curie: Paris. 180. Perdew, J.P., Chevary, J.A., Vosko, S.H., Jackson, K.A., Pederson, M.R., Singh, D.J., and Fiolhais, C., Atoms, molecules, solids, and surfaces: Applications of the generalized gradient approximation for exchange and correlation. Physical Review B, 1992. 46(11): p. 6671-6687. 181. Payne, M.C., Teter, M.P., Allan, D.C., Arias, T.A., and Joannopoulos, J.D., Iterative minimization techniques for ab initio total-energy calculations: molecular dynamics and conjugate gradients. Reviews of Modern Physics, 1992. 64(4): p. 1045. 182. Monkhorst, H.J. and Pack, J.D., Special points for Brillouin-zone integrations. Phys. Rev. B, 1976. 13: p. 5188-5192. 183. Nilekar, A.U. and Mavrikakis, M., Improved oxygen reduction reactivity of platinum monolayers on transition metal surfaces. Surface Science, 2008. 602(14): p. L89-L94.

184. Xu, Y., Ruban, A.V., and Mavrikakis, M., Adsorption and Dissociation of O2 on Pt−Co and Pt−Fe Alloys. Journal of the American Chemical Society, 2004. 126(14): p. 4717-4725. 185. Helkeman, G., Uberuaga, B.P., and Jonsson, H., A climbing image nudged elastic band method for finding saddle points and minimum energy paths. J. Chem. Phys., 2000. 113(22): p. 9901. 186. Bard, A.J. and Faulkner, L.R., Electrochemical Methods Fundamentals and Applications. Second edition ed2001: John Wiley and Sons, inc. 187. Tsiplakides, D. and Vayenas, C.G., The absolute potential scale in solid state electrochemistry. Solid State Ionics, 2002. 152–153(0): p. 625-639.

266 References

188. Burgess, I., Zamlynny, V., Szymanski, G., Lipkowski, J., Majewski, J., Smith, G., Satija, S., and Ivkov, R., Electrochemical and Neutron Reflectivity Characterization of Dodecyl Sulfate Adsorption and Aggregation at the Gold−Water Interface. Langmuir, 2001. 17(11): p. 3355-3367. 189. Besio, G.J., Prud'homme, R.K., and Benziger, J.B., Ellipsometric observation of the adsorption of sodium dodecyl sulfate. Langmuir, 1988. 4(1): p. 140-144. 190. Chatenet, M., Molina-Concha, M.B., El-Kissi, N., Parrour, G., and Diard, J.P., Direct rotating ring-disk measurement of the sodium borohydride diffusion coefficient in sodium hydroxide solutions. Electrochim. Acta, 2009. 54(18): p. 4426-4435. 191. Concha, B.M., Chatenet, M., Coutanceau, C., and Hahn, F., In situ infrared (FTIR) study of the borohydride oxidation reaction. Electrochemistry Communications, 2009. 11(1): p. 223-226. 192. Domínguez, H., Structural Transition of the Sodium Dodecyl Sulfate (SDS) Surfactant Induced by Changes in Surfactant Concentrations. The Journal of Physical Chemistry B, 2011. 115(43): p. 12422-12428. 193. Santos, D.M.F. and Sequeira, C.A.C., Determination of Kinetic and Diffusional Parameters for Sodium Borohydride Oxidation on Gold Electrodes. Journal of the Electrochemical Society, 2009. 156(5): p. F67-F74. 194. http://www.intelligent-energy.com/. 2011 [cited 2013 06/2013]. 195. http://automobiles.honda.com/fcx-clarity/. [cited 2013 12/2013]. 196. http://www.ballard.com/fuel-cell-applications/bus.aspx. [cited 2014 1/2014]. 197. Wang, K., Lu, J., and Zhuang, L., Direct determination of diffusion coefficient for borohydride anions in alkaline solutions using chronoamperometry with spherical Au electrodes. Journal of Electroanalytical Chemistry, 2005. 585(2): p. 191-196. 198. Cheng, H. and Scott, K., Determination of kinetic parameters for borohydride oxidation on a rotating Au disk electrode. Electrochimica Acta, 2006. 51(17): p. 3429-3433. 199. Santos, D.M.F. and Sequeira, C.A.C., Determination of Kinetic and Diffusional Parameters for Sodium Borohydride Oxidation on Gold Electrodes. J. Electrochem. Soc., 2009. 156(5): p. F67-F74.

267 References

200. Santos, D.M.F. and Sequeira, C.A.C., Cyclic voltammetry investigation of borohydride oxidation at a gold electrode. Electrochimica Acta, 2010. 55(22): p. 6775-6781. 201. Pletcher, D., A First Course in Electrode Processes. second edition ed2009: RSC Publishing. 202. Merino-Jimenez, I., Ponce de Leon, C., and Walsh, F.C., The use of surfactants to suppress hydrolysis of borohydride in direct borohydride fuel cells. Submitted to the J. Power Sources, Jun, 2013. 203. Khan, A.M. and Shah, S.S., Determination of Critical Micelle Concentration (CMC) of Sodium Dodecyl Sulfate (SDS) and the Effect of Low Concentration of Pyrene on its Cmc Using ORIGIN Software. J. Chem. Soc. Pak., 2008. 30(2): p. 186-191. 204. Rahman, A. and Brown, C.W., Effect of pH on the critical micelle concentration of sodium dodecyl sulphate. J. Appl. Polym. Sci., 1983. 28(4): p. 1331-1334. 205. Friedrich, J.M., Ponce-de-León, C., Reade, G.W., and Walsh, F.C., Reticulated vitreous carbon as an electrode material. J. Electroanal. Chem., 2004. 561(0): p. 203-217. 206. Tentorio, A. and Casolo-Ginelli, U., Characterization of reticulate, three — dimensional electrodes. J. Applied Electrochem., 1978. 8(3): p. 195-205. 207. Aardahl, C.L. and Rassat, S.D., Overview of systems considerations for on- board chemical hydrogen storage. International Journal of Hydrogen Energy, 2009. 34(16): p. 6676-6683. 208. Ponce de León, C. and Pletcher, D., Removal of formaldehyde from aqueous solutions via oxygen reduction using a reticulated vitreous carbon cathode. J. Appl. Electrochem., 1995. 25: p. 307-314. 209. Gyenge, E., Jung, J., and Mahato, B., Electroplated reticulated vitreous carbon current collectors for lead–acid batteries: opportunities and challenges. Journal of Power Sources, 2003. 113(2): p. 388-395. 210. Finkelstein, D.A., Da Mota, N., Cohen, J.L., and Abruna, H.D., Rotating Disk - - Electrode (RDE) Investigation of BH4 and BH3OH Electro-oxidation at Pt and - Au: Implications for BH4 Fuel Cells. Journal of Physical Chemistry C, 2009. 113(45): p. 19700-19712.

268 References

211. G. Denuault, M.V.M., A.J. Bard, Journal of Electroanalytical Chemistry, 1991. 308: p. 27. 212. Gardiner, J.A. and Collat, J.W., Polarography of the Tetrahydroborate Ion. The Effect of Hydrolysis on the System. Inorganic Chemistry, 1965. 4(8): p. 1208- 1212. 213. Santos, D.M. and Sequeira, C.A., Cyclic Voltammetry Investigation of Borohydride Oxidation at a Gold Electrode. ECS Meeting Abstracts, 2009. 902(10): p. 1053. 214. Demirci, B., Akdim, O., Andrieux, J., Hannauer, J., Chamoun, R., and Miele, P., Sodium Borohydride Hydrolysis as Hydrogen Generator: Issues, States of the Art and Applicability Upstream from a Fuel Cell. Fuel Cells, 2010. 10(3): p. 335-350. 215. Luo, X.-j., Xu, Y.-h., and Zhang, Z.-x., Study on application of platinum-coated titanium electrode to electrolysis functional water producing device. Diandu Yu Tushi, 2009. 28(Copyright (C) 2013 American Chemical Society (ACS). All Rights Reserved.): p. 72-74. 216. Sakalis, A., Fytianos, K., Nickel, U., and Voulgaropoulos, A., A comparative study of platinised titanium and niobe/synthetic diamond as anodes in the electrochemical treatment of textile wastewater. Chemical Engineering Journal, 2006. 119(2–3): p. 127-133. 217. www.water.org.uk. [4-11-2013]. 218. Xu, D., Zhang, H., and Ye, W., Hydrogen generation from hydrolysis of alkaline sodium borohydride solution using Pt/C catalyst. Catal Commun, 2007. 8: p. 1767-1771. 219. Demirci, U.B. and Garin, F., Promoted sulphated-zirconia catalysed hydrolysis of sodium tetrahydroborate. Catal Commun, 2008. 9(6): p. 1167-1172. 220. Simagina, V., Storozhenko, P., Netskina, O., Komova, O., Odegova, G., Samoilenko, T., and Gentsler, A., Effect of the nature of the active component and support on the activity of catalysts for the hydrolysis of sodium borohydride. Kinet Catal, 2007. 48(1): p. 168-175.

221. Khain, V., Some characteristic features of redox reactions involving the BH4 ions in an aqueous medium. Russ J Inorg Chem, 1983. 28: p. 1410-1413.

269 References

222. Corrias, A., Ennas, G., Musinu, A., Marongiu, G., and Paschina, G., Amorphous transition metal-boron ultrafine particles prepared by chemical methods. Chem Mater, 1993. 5(12): p. 1722-1726. 223. Lee, J., Kong, K.Y., Jung, C.R., Cho, E., Yoon, S.P., Han, J., Lee, T.-G., and

Nam, S.W., A structured Co–B catalyst for hydrogen extraction from NaBH4 solution. Catal Today, 2007. 120(3-4): p. 305-310. 224. Krishnan, P., Advani, S.G., and Prasad, A.K., Thin-film CoB catalyst templates for the hydrolysis of NaBH4 solution for hydrogen generation. Appl Catal, B, 2009. 86(3–4): p. 137-144. 225. Kim, D.-R., Cho, K.-W., Choi, Y.-I., and Park, C.-J., Fabrication of porous Co– Ni–P catalysts by electrodeposition and their catalytic characteristics for the

generation of hydrogen from an alkaline NaBH4 solution. Int J Hydrogen Energy, 2009. 34(6): p. 2622-2630. 226. Akdim, O., Demirci, U.B., and Miele, P., Deactivation and reactivation of cobalt in hydrolysis of sodium borohydride. Int J Hydrogen Energy, 2011. 36(Copyright (C) 2012 American Chemical Society (ACS). All Rights Reserved.): p. 13669-13675. 227. Gregg, S.J. and Sing, K.S.W., Adsorption, surface area, and porosity. 2nd ed1982, London: Academic Press. 228. Streletzky, K. and Phillies, G.D.J., Temperature Dependence of Triton X-100 Micelle Size and Hydration. Langmuir, 1995. 11(1): p. 42-47. 229. Gonick, E. and McBain, J.W., Cryoscopic Evidence for Micellar Association in Aqueous Solutions of Non-ionic Detergents. J. Am. Chem. Soc., 1947. 69(2): p. 334-336. 230. Szymczyk, K., The Properties of Binary Mixtures of Ethoxylated Octyl Phenols with Ethoxylated Fluorinated Alkanols at the Water/Air Interface. Journal of Surfactants and Detergents, 2011. 14(3): p. 415-423. 231. http://www.chemicalbook.com/ChemicalProductProperty_EN_CB2147453.htm. 2012 11/02/2012]. 232. Brezinski, D., Laying the Foundation for New Technologies. Paint and coating industry (PCI), 2003. 19(1): p. 36.

270